Fidatezza PDF
Fidatezza PDF
1
Dependability: Taxonomy
reliability
availability
measures maintainability
safety
security
fault forecasting
dependability means fault tolerance
fault removal
fault prevention
faults
threats errors
failures
2
Quantitative analysis
The quantitative analysis aims at numerically
evaluating measures to characterize the dependability
of an item:
3
Design specifications
Technological items must be dependable.
4
Technical assistance and maintenance
The planning of all the activity related to the technical
assistance and maintenance is linked to the system
dependability (expected number of failure in time).
5
Life Cycle Cost
6
Market competition
7
Purpose of evalua-on
• Understanding a system
– Observation
– Operational environment
– Reasoning
Model-Based
ü Less believable, Less expensive
ü Analytic vs Discrete-Event Simulation
ü Combinatorial vs State-Space Methods
9
Measurement-Based
Most believable, most expensive;
Data are obtained observing the behavior of physical
objects.
v field observations;
v measurements on prototypes;
v measurements on components (accelerated tests).
10
Methods of evalua-on
• Measurements + Models data bank
11
The probabilistic approach
The mechanisms that lead to failure a technological
object are very complex and depend on many physical,
chemical, technical, human, environmental … factors.
12
Reliability
13
Basic Defini-ons: cdf
• Let X be the random variable represen-ng the -me
to failure of an item.
F(t) = Pr { X ≤ t }
F(t) represents the probability that the item is
already failed at time t (unreliability) .
14
Basic Defini-ons: cdf
15
Basic Defini-ons: cdf
F(t)
1
F(b)
F(a)
0 a b t
F(0) = 0
lim F(t) = 1
tà∞
F(t) = non-decreasing
16
Basic Defini-ons: Reliability
• Let X be the random variable represen-ng the -me
to failure of an item.
17
Basic Defini-ons
Ø Reliability
Ø Survivor func-on
18
Basic Defini-ons: Reliability
R(t)
1
R(a)
0 a b t
R(0) = 1
lim R(t) = 0
tà∞
R(t) = non-increasing
19
Basic Defini-ons: density
• Let X be the random variable represen-ng the -me
to failure of an item and let F(t) be a derivable cdf:
d F(t)
f (t) = ———
dt
f (t) dt = Pr { t ≤ X < t + dt }
20
Basic Defini-ons: Density
f (t)
0
a b t
b
∫ f(x) dx = Pr { a < X ≤ b } = F(b) – F(a)
a
t ∞
∫ f(x) dx = F(t) R(t)=1- F(t)=∫ f(x) dx
0 t 21
Basic Defini-ons: DensityàExpected Life
f (t)
1
0
t
∞ ∞
MTTF = E [t ] = ∫ 0
tf (t )dt = ∫ 0
R (t ) dt
Integration by parts àsee demonstration 22
01/11/16 23
Basic Defini-ons
Ø density func-on
t
F (t ) = ∫ f ( x)dx
Ø density dF −∞
t
dt = ∫ f ( x)dx ,
Ø f(t) = 0
For a non-negative
random variable
24
Hazard (failure) rate
f (t ) f (t )
h (t ) = =
R(t ) 1 − F (t )
25
The Failure Rate of a Distribu-on
()
• h t Δ t is the conditional probability that the unit
will fail in the interval ( t , t + Δ t ) given that it is
functioning at time t.
26
h(t) Bathtub curve
CFR
Constant fail. rate
(useful life)
DFR IFR
28
Useful life (cfr)
The failure rate remains constant in time (age
independent) .
29
Wear-out phase (ifr)
The failure rate increases with age.
30
Exponen-al Distribu-on
Failure rate is age-independent (constant).
F(t)
1.0
0.5
F(t) = 1 - e - λ t
32
The Reliability Func-on of an Exponen-ally
Distributed Random Variable With Parameter λ = 1
R(t)
1.0
R(t) = e - λ t
0.5
33
Exponen-al Density Func-on (pdf)
f(t)
MTTF = 1/ λ
34
Memoryless Property of the
Exponen-al Distribu-on
• Assume X > t. We have observed that the
component has not failed until time t
• Let Y = X - t , the remaining (residual) lifetime
Gt ( y ) = P (Y ≤ y | X > t )
= P( X ≤ y + t | X > t )
P (t < X ≤ y + t )
= = 1 − e − λy
P( X > t )
35
Memoryless Property of the
Exponen-al Distribu-on (cont.)
36
Quiz 3:
If two components (say, A and B) have independent iden-cal
37
Weibull Distribu-on
− λ tα
• Distribu-on Func-on: F (t ) = 1 − e t≥0
• Density Func-on: α −1 − λ t α
f (t ) = λα t e t≥0
• Reliability: −λ tα
R(t ) = e t≥0
38
Weibull Distribu-on
α : shape parameter;
λ : scale parameter.
f (t ) α −1
Failure Rate: h(t ) = = λα t t≥0
R (t )
α <1 Dfr
α =1 Cfr
α >1 Ifr
39
Failure Rate of the Weibull Distribu-on with Various
Values of α
40
Weibull Distribu-on for Various
Values of α
Cdf density
Consider a component represented by a Weibull distribution with
shape parameter, α=4 and scale parameter, λ=1000. Calculate its
reliability at an operating time of 1500 h. 41
Failure Rate Models (cont.)
• There are several ways to incorporate time dependent
failure rates in availability models
• The easiest way is to approximate a continuous function
by a piecewise constant step function
7 Discrete Failure-Rate Model
Failure-Rate Multiplier
6 λ1
5
λ2 λ SS
4
43
A life-me experiment
X1
1
X2
2
X3
3
X4
4
XN
N
t=0
N XN
45
Interpre-ng reliability tests
CUMULATIVE TIME
The real MTBF will be obtained only when all the components will
be failed
Time truncated tests
Failure truncated Time truncated
n=6 n=8
α=0,9 è χ2=2,20èMTBF=90909h α=0,9 è χ2=3,49èMTBF=57306h
α=0,1 è χ2=10,6èMTBF=18868h α=0,1 è χ2=13,4èMTBF=14925h
54
Exercise 1
3. Compare the point value obtained for the (a) case and assign it a confidence level.
4. Calculate the limit MTBFs in case the 90%confidence level should be double-sided.
Weibull chart
57
58
Exercise 3