Quiz 1

Queensland University of Technology
Semester 1 2016
ENB432: Engineering Asset Management and
Maintenance Laurisa Swamy;N8888582
QUIZ 1
Question 1.
a)
By assuming that pump failures occur at random and can be modelled by a
Poisson distribution, the average rate of occurrence of random events, ,
is
required to calculate probability. Considering that this type of pump has one
failure in three years.
Therefore,
1
= =0.3
3
Let T = time between failure
P (T >4 )=1P (T 4 )
Using excel this equates to,
1EXPONDIST ( 4, 0.333333,1 )=0.263597
b)
The probability that a completely new pump will operate for 4 years is calculated
below.
1
= =0.3
3
P (T >4 )=1P (T 4 ) Using excel this equates to,
1EXPONDIST ( 4, 0.333333,1 )=0.263597
c)
As the failure rate follows a Poisson distribution this assumes that, T (time
between failures) follows an exponential distribution. As such the probability of
the pump operating for another four years after two years of operation can be
calculated using Markovs property of exponential distribution.
T x
T x+t |T > t )=P( )
P
The conditional probability distribution depends only upon present state and
does not account for preceding events when calculating probability. This is
referred to as memoryless property.
Therefore, the probability that the pump will continue operating for 2+4 years is:
T 2+ 4T > 2
T > 4+ 2|T >2 )=1P()
P
1P(T 4)
As previously stated,
=0.3
Using excel this equates to,
1EXPONDIST ( 4, 0.333333,1 )=0.263597
Question 2.
a)
For constant failure rates, the function becomes
( t )= . It is known that the
hazard rate is estimated to be 1 over 500 hours (constant).

Therefore, the hazard rate,
( t )=
1
=.002= .MTTF, can be expressed as:
500
= tf ( t ) dt
0
Where f(t) is the failure density function
For constant failure rates, the failure density function can be expressed as:
f ( t )=
dF ( t )
dt
d ( 1et )
dt
et
As a result, MTTF becomes:
= tf ( t ) dt
0
t e t dt
0
Substituting in numbers,
1
0.002
500 hours
Therefore, the MTTF for the sample of failure times of a light bulb is 500 hours.
b)
Given that the light bulb has a constant failure rate, it indicates an exponential
distribution for the failure density function. As such, the memoryless property of
Markovs property of exponential distribution can be applied.
Therefore, the probability of a light bulb operating for another 200 hours after
operating for 400 hours is:
T 200
T 200+400|T >400 )=P()
P
Expanded further,
P (T 200 )=F ( t=200 )

From part a)
=0.002
Thus,
0.002200
0.67032
c)
For Type 1 censored data the confidence interval of
is:
x 22r ,(1 /2)
^=1
^(x 22r , /2 )
2r
Pr
where r is the number of recorded failures
2r is the degrees of freedom for the chi-square random variable
is the pre-determined level of significance.
This equation can be used to calculate 95% confidence intervals for the hazard
rate and the MTTF.
Expanded from the above equation the lower and upper bounds of the
confidence intervals are calculated using:
Lower bound :
^(x22 r , / 2)
2r
x22 r ,(1 /2)
^
Upper bound :
The formulas used in excel for Chi square distribution are:
Lower limit :CHIINV (1 ,2r )

2
Upper limit :CHIINV ( , 2r)

2
From the sample size of 80, 30 are right censored. Therefore, the recorded
number of failures, r = 50. The constant failure rate is estimated to be
1/500/hour and a 95% confidence interval gives an
0.05
1
, 250 )
(
2
Lower bound=CHIINV
value of 0.05
0.001484
( 250 )( 0.002 )
Upper bound=CHIINV
( 0.052 , 250)
( 250 )( 0.002 )
0.002591 The confidence interval of the
MTTF for a constant failure rate is calculated by taking the reciprocal of the
confidence interval limits of
Therefore, the 95% confidence interval of MTTF is:
Lower=
Upper=
1
0.002591
1
0.001484
385.918
637.6554
From these results it can be concluded that 95% of the time over repeated
sampling, the constant hazard rate is estimated to be within the interval [0.0015,
0.0026]. Additionally, for 95% of the time over repeated sampling, the MTTF is
estimated to be within the interval [385.918, 637.6554].
Question 3.
a)
To fit the Weibull Distribution, the data is sorted in ascending order, shown below.
Failure Time
Data
108
17
7
26
24
20
51
29
76
31
12
44
25
56
36
Sorted
Data
7
12
17
20
24
25
26
29
31
36
44
51
56
76
108
Next the mean order numbers (k) are assigned, these are also used to calculate
cumulative distribution function estimations (ECDF). The excel formula used to
calculated ECDF is BETAINV (0.5, k, n-k+1) where n is the sample number. This
step is show below.
Failure Time
Data
Sorted
Data
108
17
12
17
26
20
24
24
20
25
51
29
26
29
7
8
76
31
31
36
10
12
44
11
44
51
12
25
56
56
76
13
14
ECDF
(k)
0.0451
58
0.1093
96
0.1743
21
0.2393
93
0.3045
2
0.3696
7
0.4348
33
0.5
0.5651
67
0.6303
3
0.6954
8
0.7606
07
0.8256
79
0.8906
36
108
15
04
0.9548
42
The parameters of the Weibull distribution can now be determined. This is done
by transforming the data by X = log(xdata) and Y = log(-log(1-ECDF)). This
function in excel and the outputs can be seen below.
Failure Time
Data
Sorted
Data
ECDF
(k)
X
1.9459
1
108
0.0451
58
17
12
0.1093
96
2.4849
07
17
0.1743
21
2.8332
13
2.9957
32
26
20
0.2393
93
24
24
0.3045
2
3.1780
54
20
25
0.3696
7
3.2188
76
51
26
0.4348
33
3.2580
97
29
29
0.5
3.3672
96
76
31
0.5651
67
3.4339
87
31
36
10
12
44
11
44
51
12
25
56
13
56
76
14
36
108
15
0.6303
3
0.6954
8
0.7606
07
0.8256
79
0.8906
04
0.9548
42
3.5835
19
3.7841
9
3.9318
26
4.0253
52
4.3307
33
4.6821
31
Y
3.0745631
22
2.1554114
66
1.6526126
26
1.2959455
71
1.0129317
33
0.7732461
33
0.5610078
9
0.3665129
21
0.1829693
01
0.0048684
9
0.1731287
13
0.3574276
2
0.5578191
51
0.7942496
5
1.1306208
57
Then a linear line is fitted to X and Y by either regression or a scattered plot with
a line of best fit. Both of these methods are shown below.
As can be seen, both outputs tell us that that the shape parameter
and x-
intercept are 1.6183 and

-6.046 respectively. The scale parameter
can now be computed using the
following excel function:
=exp
( b )=41.93
07
b)
Question 4.
a)
Rs= R1 X R2 X R3
R1= (1-(1-R)(1-R))
R2=(1-(1-R*R)*(1-R))
R3=(3R^2-2R^3)
Rs=(1-(1-R)(1-R))* (1-(1-R*R)*(1-R))* (3R^2-2R^3)
= -2R^8+9R^7-11R^6-R^5+6R^4
b)

Quiz 1

Uploaded by

Copyright:

Available Formats

Quiz 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quiz 1

Uploaded by

Copyright:

Available Formats

Queensland University of Technology

Using excel this equates to,

1EXPONDIST ( 4, 0.333333,1 )=0.263597

1EXPONDIST ( 4, 0.333333,1 )=0.263597

( t )= . It is known that the

hazard rate is estimated to be 1 over 500 hours (constant).

Where f(t) is the failure density function

As a result, MTTF becomes:

P (T 200 )=F ( t=200 )

x 22r ,(1 /2)

is the pre-determined level of significance.

x22 r ,(1 /2)

Lower limit :CHIINV (1 ,2r )

Upper limit :CHIINV ( , 2r)

0.002591 The confidence interval of the

Therefore, the 95% confidence interval of MTTF is:

intercept are 1.6183 and

can now be computed using the

following excel function:

You might also like