Chapter 4
Time-Dependent Failure Rate Models
Part I
The Weibull Distribution
Minimum Extreme Value Distribution
Chapter 4 1
Time Dependent Failure Mode
One of the more useful probability distributions in reliability is the
Weibull. The Weibull failure distribution may be used to model both
increasing and decreasing failure rates.
Assume l(t) is a power function: l(t ) = at ; a > 0
b
rewrite as:
where
b is the shape parameter and q is a scale parameter
Chapter 4 2
Reliability Function
b
b -1
é t b æ t¢ ö ù ætö
-ç ÷
êë
ò
R(t ) = exp ê - R
0 q çq ÷
eliability
è ø
dt ¢ú = e
úû
èq ø
1.2
R(t) Beta
1 0.5 1.5
0.8 2.0 4.0
0.6
Theta = 2
0.4
0.2
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
t
-0.2
Chapter 4 3
The Probability Density Function (PDF)
1.0
f(t)
0.9
0.8
Beta
0.5
0.7 1.5
2.0
0.6
4.0
0.5
0.4 Theta = 2
0.3
0.2
0.1
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
t
-0.1
b -1
d R (t ) b æ t ö -( t q )
b
f (t ) = - = ç ÷ e
dt q èq ø
Chapter 4 4
Weibull Shape Parameter - b
Value Property
0<b<1 Decreasing Failure Rate (DFR)
b=1 Exponential Distribution (CFR)
1<b<2 IFR-concave
b=2 Rayleigh Distribution (LFR)
b>2 IFR - Convex
3 <= b <= 4 IFR - Approaches Normal
Distribution - Symmetrical
• (LFR-Linear Failure Rate)
b = 3.43927 Most closely approximates the normal
b = 3.43938 Mean = median
Chapter 4 5
The Mean Time to Failure (MTTF)
æ 1ö
MTTF = q G ç1 + ÷
è bø
¥
G (x) = the gamma function ò e dy
x -1 - y
G ( x ) = y
0
G( x) = ( x - 1)G( x - 1)
e.g. G ( 4.23) = ( 3.23) G ( 3.23) = ( 3.23)( 2.23) G ( 2.23)
= ( 3.23)( 2.23)(1.12023) = 8.0689
1
lim MTTF = lim q G (1 + ) = qG (1) = q
Chapter 4
b ®¥ b ®¥ b 6
Gamma Function - selected values
x Gamma(x) x Gamma(x)
1.53 .88757 2.23 1.12023
1.54 .88818 2.24 1.12657
1.55 .88887
2.25 1.133
1.56 .88964
1.57 .89049 2.26 1.13954
1.58 .89142 2.27 1.14618
1.59 .89243 2.28 1.15292
1.6 .89352 2.29 1.15976
1.61 .89468 2.3 1.16671
1.62 .89592 2.31 1.17377
1.63 .89724
2.32 1.18093
1.64 .89864
2.33 1.18819
1.67 .9033
1.68 .905 2.34 1.19557
1.69 .90678 2.35 1.20305
Appendix Table A.9
Chapter 4 7
æ 1ö
MTTF = q G ç1 + ÷
è bø
Chapter 4 8
Chapter 4 9
Example Problems
Let T = a random variable, the time Let T = a random variable, the time
to failure of a circuit card T has a to failure of a fuse.
Weibull distribution with beta equal T has a Weibull distribution with beta
to 0.5 and theta equal to 500 equal to 1.5 and theta equal to 500
(thousands of hours). (thousands of hours).
Find R(50,000) and the MTTF. Find R(50,000) and the MTTF.
æ 1ö
MTTF = q G ç1 + ÷
Chapter 4
è bø 10
The Variance
ìï æ 2 ö é æ 1 ö ù
2
üï
s 2 = q 2 íG ç 1 + ÷ - ê G ç 1 + ÷ ú ý
ïî è b ø ë è b ø û ïþ
s
lim = 0
2
b ®¥
Chapter 4 11
Example Problem - standard deviation
s 2 = 5002 [G (5) - 22 ] s 2 = 5002 [G (1 + 4 / 3) -.9032 ]
= 500 [24 - 4] = 5,000,000
2
= 5002 [118819
. -.81595]
s = 2236 (thousands of hr.) = 93059.8
s = 305 (thousands of hr.)
Chapter 4 12
Design Life and Median
b
-( t q )
set R(t ) = e =R
1b
Solving for t: t R = q ( - ln R )
t0.50 = tmed = q (- ln 0.5)1 b = q (0.69315)1 b
Chapter 4 13
The Mode
b -1
bætö -( t q )
b
max f (t ) = ç ÷ e
t ³0 q èq ø
b -2 2 b -2
df (t ) b ( b - 1) æ t ö -( t q )
b b 2
ætö -( t q )
b
= e - 2 e =0
dt q 2 çè q ÷ø q ç ÷
èq ø
Solve for t:
ìq (1 - 1 b )1/ b for b > 1
tmode =í
î0 for b £ 1
Chapter 4 14
Example - design life, median, & mode
beta = .5 beta = 1.5 = 3/2
t.9 = 500 (-ln.90)2 t.9 = 500 (-ln.90)2/3
= 5.55 (1000 hr.) = 111.5 (1000 hr.)
tmed = 500 (.69315)2 tmed = 500 (.69315)2/3
= 240 (1000 hr.) = 391.6 (1000 hr.)
tmode = 0 tmode = 500 [1-2/3)]2/3
= 240 (1000 hr.)
Chapter 4 15
Conditional Reliability # Burn-In
R ( t + T0 )
R(t T0 ) =
R (T0 )
b b
b
exp{-[(t + T0 ) q ] } é æ t + T0 ö æ T0 ö ù
R(t T0 ) = b
= exp ê - ç ÷ +ç ÷ ú
exp[-(T0 q ) ] êë è q ø è q ø úû
Chapter 4 16
Example - conditional reliability
Beta = .5; R(50) = .7289 Beta = 1.5; R(50) = .969
R(50|50) = R(100)/R(50) R(50|50) = R(100)/R(50)
=exp [-(100/500).5] / .7289 =exp [-(100/500).1.5] / .969
= .6394 / .7289 = .8772 = .9144 / .969 = .9437
Chapter 4 17
EXAMPLE 4.4
An electrical system has four series connectors each
having a Weibull failure law with b = 3/4 and θ = 2000
hr. Find the reliability of the system of four connectors
for 150 hours.
é æ 150 ö3 4 ù
R(150) = exp ê-4 ç ÷ ú = 0.5637
êë è 2000 ø úû
Chapter 4 18
Three Parameter Weibull
t ¢ = t - t0 .
Let t0 be the minimum life such that T>t0, then
b -1
b æ t - t0 ö
l (t ) = ç t ³ t0
q è q ÷ø
é æ t - t0 ö b ù
R(t ) = exp ê - ç ÷ ú t ³ t0
êë è q ø úû
æ 1ö 1b
MTTF = t0 + q G ç1 + ÷ tmed = t0 + q (0.69315)
è bø
Transform the three-parameter Weibull into the two-parameter
Weibull with the transformation t’ = t – t0 if t0 is known!
Chapter 4 19
Redundancy - Two Identical Weibull
Components
2
R s (t) = 1 - 1 - R(t )
MTTF
b -1 - ( t q )b
bætö 2 - 2e
ls (t ) = ç ÷
q èq ø 2-e - ( t q )b
Not a Weibull hazard rate function. but as t increases, ls(t) approaches l(t).
Chapter 4 20
Student Exercise - Weibull
A certain brand lightning arrester has a Weibull failure
distribution with a shape parameter of 2.4 and a characteristic
life of 10 years. Find:
a. R(5 yrs)
b. MTTF
c. Standard deviation
d. Median and Mode
e. 99% (B1) and 95% B(5) design life
f. R(5|5)
Chapter 4 21
Student Exercise - solution
F
G5I
J
2 .4
a. R(5) = e H
10 K
-
=.8274
b. MTTF = 10 G (1 + 1 / 2.4) = 10 G (142
. )
10 (.88636) = 8.86 yrs
m
c. s 2 = 102 G (1 + 2 / 2.4) -.886362 r
or s = 3.93 yr.
Chapter 4 22
Student Exercise - solution (continued)
d . tmed = 10 .69315 b g = 8.58 yr.
1/ 2.4
= 10b 1 - 1 / 2.4g = 8.0 yr.
1/ 2 .4
t mod e
= 10b- ln.99g = 15
1/ 2 .4
e. t.99 . yr.
= 10 b
- ln.95g = 2.9 yr.
1/ 2 .4
t.95
Chapter 4 23
Student Exercise - solution
(continued)
2.4
æ 10 ö
-ç ÷
è 10 ø
R(10) e
f . R(5 | 5) = =
R(5) .8274
.3679
= = .4446
.8274
Chapter 4 24