= http://www.rt.cs.boeing.com/MEA/stat/reliability.html
i=1 i i=1 i
t
= exp
regardless of the value > 0
Z
with (t) = 0 exp(x) xt1 dx
The variance of T
Z
= E(T ) =
2 2
0 (t ) f (t) dt = (1 + 2/) (1 + 1/)
2 2 2
P (t T t + dt) f (t) dt
Z
t
F (t) = 0 f (x) dx
p
cumulative distribution function
0
=0 = 1000 = 2000
= 1000, = 2.5
probability density
0 2000 4000
cycles
= 1000
= 0, = 2.5
probability density
= 2000
= 3000
cycles
=7
= .5
probability density
=4
=1
=2
= 0, = 1000
cycles
Z ! Z !
t t
F (t) = 1exp 0 (x) dx and f (t) = (t) exp 0 (x) dx
63.2% 36.8%
characteristic life = 30,000
shape = 2.5
0.00002
true model
estimated model from 9 data points
0.0
cycles
shape parameter = 4
63.2% 36.8%
p=P(T < t )
t =t p p-quantile
25390 3.02 33860 4.27 29410 5.01
15000
10000
tp
5000
Note that
log (tp) = log() + wp/ , where wp = log [ log(1 p)]
Expect points log[T(i)], wpi , i = 1, . . . , n, to cluster around line
with slope 1/ and intercept log()
f, (t1)dt f, (tn)dt
.990
.900
.632
.500
.200
probability
.100
.010 true model, Weibull( 100 , 3 )
m.l.e. model, Weibull( 108 , 3.338 )
least squares model, Weibull( 107.5 , 3.684 )
.001
cycles/hours
.990
.900
.632
.500
.200
probability
.100
.010 true model, Weibull( 100 , 3 )
m.l.e. model, Weibull( 102.8 , 2.981 )
least squares model, Weibull( 102.2 , 3.144 )
.001
cycles/hours
Let F,
b b (t) be the fitted Weibull distribution function
#{Ti t; i=1,...,n}
Let F
d
n (t) = n be the empirical distribution function
Compute a discrepancy metric D between F,
b
d
b and Fn ,
b , Fn ) = sup F b b (t) Fn (t)
d d
DKS (F,
b ,
Kolmogorov-Smirnov
t
Z !
2
b (t) Fn (t)
d d
DCvM (F,
b b , Fn ) =
0 F,
b f,
b b (t) dt Cramer-von Mises
!
2
b (t) Fn (t)
d
d
Z
F,b
DAD (F,
b b , Fn ) = f,
b b (t) dt Anderson-Darling
0
F,
b b (t)(1 F b
b
, (t))
The distributions of D, when sampling from a Weibull population,
are known and p-values of observed values d of D can be calculated
p = P (D d) = BCSLIB: HSPFIT
1.0
n = 10
0.8
K-S distance
0.6
0.4
0.2
0.0
1
p(CvM) = 0.54 p(CvM) = 0.27 p(CvM) = 0.62 p(CvM) = 0.64
p(AD) = 0.62 p(AD) = 0.35 p(AD) = 0.66 p(AD) = 0.69
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
n = 10
3.8 4.2 4.6 5.0 4.0 4.4 4.8 3.8 4.4 5.0 3.8 4.2 4.6 5.0
1
p(CvM) = 0.53 p(CvM) = 0.21 p(CvM) = 0.23 p(CvM) = 0.56
p(AD) = 0.6 p(AD) = 0.32 p(AD) = 0.22 p(AD) = 0.57
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
4.0 4.4 4.8 3.8 4.2 4.6 4.0 4.4 4.8 3.8 4.2 4.6
1
p(CvM) = 0.49 p(CvM) = 0.027 p(CvM) = 0.053 p(CvM) = 0.39
p(AD) = 0.49 p(AD) = 0.028
p(AD) = 0.048
p(AD) = 0.38
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
n = 20
3.5 4.0 4.5 5.0 2.5 3.5 4.5 4.2 4.6 3.8 4.2 4.6 5.0
1
p(CvM) = 0.41 p(CvM) = 0.63 p(CvM) = 0.6 p(CvM) = 0.017
p(AD) = 0.41 p(AD) = 0.69
p(AD) = 0.64
p(AD) = 0.023
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
3.6 4.0 4.4 4.8 3.0 4.0 5.0 4.0 4.4 4.8 3.6 4.0 4.4 4.8
1
p(AD) = 0.43
p(AD) = 0.32
p(AD) = 0.12
p(AD) = 0.52
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
-4
-4
-4
-4
n = 50
2.0 3.0 4.0 5.0 3.0 4.0 5.0 3.5 4.0 4.5 3.5 4.5
1
p(AD) = 0.72
p(AD) = 0.23
p(AD) = 0.62
p(AD) = 0.039
0
0
-1
-1
-1
-1
-2
-2
-2
-2
-3
-3
-3
-3
-4
-4
-4
-4
3.5 4.0 4.5 3.5 4.0 4.5 5.0 3.5 4.0 4.5 5.0 3.5 4.0 4.5 5.0
0
-2
-2
-2
-2
-4
-4
-4
-4
n = 100
3.0 4.0 5.0 3.5 4.5 3.0 4.0 5.0 3.5 4.5
0
-2
-2
-2
-2
-4
-4
-4
-4
3.0 4.0 5.0 3.0 4.0 5.0 3.5 4.5 2.5 3.5 4.5
For a target
= , = , = tp = [ log(1 p)]1/ , or = P, (T t)
one gets corresponding m.l.e c by replacing (, ) by (, c
c )
L: Lawless Exact Method (RAP); B: Bain Exact Method (Tables); M: MLE Approximate Method
90% confidence intervals
35000
true
25000
L BM
samples 1 2 3 4 5
8
90% confidence intervals
true
4
2
L BM
0
1 2 3 4 5
25000
95% lower confidence bounds
15000
true 10-percentile
5000
L BM
L: Lawless Exact Method (RAP); B: Bain Exact Method (Tables); M: MLE Approximate Method
0
samples 1 2 3 4 5
95% upper confidence bounds
L BM
0.20
1 2 3 4 5
35000
95% confidence intervals
25000
true
15000
7
95% confidence intervals
true
2
95% upper confidence bounds
0.15
true P(T < 10000)
0.05
0.0
true t .10
10000
5000
unit
10 X
9 ?
8 ?
7 X
6 X
5 ?
4 X
3 X
2 ?
1 ?
cycles
unit
10 ?
9 ?
8 ?
7 ?
6 X
5 X
4 X
3 X
2 X
1 X
cycles
unit
10 ?
9 ( ]
8 ?
7 ( ?
6 ( ]
5 ?
4 ( ]
3 ( ?
2 ?
1 ( ]
cycles
unit
10 ?
9 ?
8 ?
7 ?
6 X
5 X
4 X
3 X
2 ?
1 X
cycles
0.8
distribution functions
0.4
0.4
n = 10 n = 30
0.0
0.0
0 5000 15000 25000 0 5000 15000 25000
0.8
0.8
0.4
0.4
n = 50 n = 100
0.0
0.0
0.8
and true CDF
0.4
0.4
n = 10 n = 30
0.0
0.0
... ... . .
0.8
0.4
0.4
n = 50 n = 100
0.0
0.0
. . .. .. ... ....... . .. .
0.8
and true CDF
0.4
0.4
n = 10 n = 30
0.0
0.0
.. . ... . ... .
0.8
0.4
0.4
n = 50 n = 100
0.0
0.0
Kaplan-Meier CDF
0.8 Weibull mle for CDF
0.8
true CDF
0.4
0.4
n = 10 n = 30
0.0
0.0
. . . .. .. . .. . .. . . .. . . . .
0.8
0.4
0.4
n = 50 n = 100
0.0
0.0
1.0
superimposed is the true Weibull(10000,2.5)
that generated the 1,000 interval censored data cases
0.8
0.4
0.2
0.0
cycles
1.0
superimposed is the true Weibull(10000,2.5)
that generated the 10,000 interval censored data cases
0.8
0.4
0.2
0.0
cycles
0.8
8124 cases
0.6
0.4
0.2
estimated .99-quantile
0.0
days
d
and F (t) is the nonparametric MLE of F (t) for the censored sample
.990
.900
.632
.500
.200
probability
.100
.001
cycles/hours
For complete & type II censored data RAP computes exact coverage
confidence bounds for , , tp, and P (T t).
WEIBREG and commercial software (Weibull++, WeibullSMITH,
common statistical packages) compute approximate confidence
bounds for above targets, based on large sample m.l.e. theory
RAP & WEIBREG are Boeing code, runs within DOS mode
of Windows95 (interface clumsy, but job gets done), contact me
Boeing has a site license for Weibull++ Version 4
in the Puget Sound area
contact David Twigg (425) 717-1221, david.twigg@pss.boeing.com
There is a Weibull++ Version 5 out
0.999
0.990
alpha = 102.4 [ 82.52 , 143.8 ] 95 %
0.900
beta = 2.524 [ 1.352 , 3.696 ] 95 %
n = 30 , r = 13
0.500
0.200
0.100
0.010
0.001
0.999
0.990
alpha = 89.79 [ 77.4 , 109.4 ] 95 %
0.900
beta = 3.674 [ 2.15 , 5.198 ] 95 %
n = 30 , r = 13
0.500
0.200
0.100
0.010
0.001
= b1 + b2Z
with
b1 = log [ (VoltU )] c log (VoltU ) , b2 = c, and Z = log (Volt)
= b1 + b2 Z
with
b1 = log [(tempU )] , b2 = , and Z = temp1
tempU
log(1) = b1 1 + b2 0 + b3 0 = log(1)
.. ..
log(n1 ) = b1 1 + b2 0 + b3 0 = log(1)
.. ..
log(n1+1) = b1 1 + b2 1 + b3 0 = log(1) + [log(2) log(1)] = log(2)
.. ..
log(n1+n2 ) = b1 1 + b2 1 + b3 0 = log(1) + [log(2) log(1)] = log(2)
.. ..
log(n1+n2+1) = b1 1 + b2 0 + b3 1 = log(1) + [log(3) log(1)] = log(3)
.. ..
log(N ) = b1 1 + b2 0 + b3 1 = log(1) + [log(3) log(1)] = log(3)
200
100
thousand cycles
50
20
10
5
Voltage
200
100
thousand cycles
50 true li
ne
mle lin
e
20
10
mle for .01-quantile
5
95% lower bound for .01-quantile
Voltage
2 TTT
L, = 2
2r+2,
This also holds for r = 0, i.e., no failures at all over exposure period
T T T = T1 + + Tn .