0% found this document useful (0 votes)
293 views

Reliability Estimation Using Minitab

This document outlines steps for generating reliability data and performing reliability analyses using Minitab software. It discusses generating both Weibull and lognormal datasets in Part 1. Part 2 covers analyzing the data using parametric distribution fitting. Several distributions are compared, including Weibull, lognormal, exponential and loglogistic. Metrics like MTTF, percentiles and goodness of fit statistics are evaluated to identify the best fitting distribution.

Uploaded by

Shiva Srt
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
293 views

Reliability Estimation Using Minitab

This document outlines steps for generating reliability data and performing reliability analyses using Minitab software. It discusses generating both Weibull and lognormal datasets in Part 1. Part 2 covers analyzing the data using parametric distribution fitting. Several distributions are compared, including Weibull, lognormal, exponential and loglogistic. Metrics like MTTF, percentiles and goodness of fit statistics are evaluated to identify the best fitting distribution.

Uploaded by

Shiva Srt
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 16

IE 7270: Reliability Estimation

Fall 2004

Reliability Estimation
Using Minitab
IE 7270 Reliability Estimation Fall 2010
GS WASSERMAN

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

Using Minitab
MINITAB
HAS TWO MAIN WINDOWS
A SESSION WINDOW and
A WORKSHEET WINDOW

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

PART ONE:
Generate Dataset
Weibull
-Step 1

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

PART ONE:
Generate Dataset
Weibull
-Step 2

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

PART ONE:
Generate Dataset
Lognormal
-Step 2

ln tmed

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

PART TWO:
Data Analysis

Most widely
Used procedure

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

Time or Failure

Gives:
-MLE or LSXY
-Confidence level

Distribution to fit:

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

STAT > REL. PARAMETRIC


RIGHT CENSORING
Probability Plot for Time
Normal Distribution - ML Estimates - 95.0% CI
Censoring Column in Censored = 0
99

95
90
80

946.19
124.11

MTTF
StDev

946.19
124.11

Median
IQR

946.19
167.43

Failure
Censor

70

Percent

Location
Scale

60
50
40

AD*

5
1
1.9995

30
20
10
5

1
500

1000

1500

Time to Failure

Probability Plot for Time


Normal Distribution - ML Estimates - 95.0% CI
Censoring Column in Censored = 0
99

95
90

Percent

80

Location
Scale

946.19
124.11

MTTF
StDev

946.19
124.11

Median
IQR

946.19
167.43

70

Failure
Censor

60
50

AD*

5
1
1.9995

40
30
20
10
5

1
500

1000

1500

Time to Failure

V13 output shown here

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

Column of Censor Flags

For multiply-censored
Data, create a column of
Indicator variables 0 or 1
-- to denote censored or not!

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

OPTIONS
Graph

Maximum Likelihood
or Rank Regression
NOTE: In Minitab V15,
the plotting position options
are found in Tools ..> Options
on the main menu. Select probability
plots, and explore the options within!

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

STAT > REL. PARAMETRIC


RIGHT CENSORING

Distribution Analysis: Time


Variable: Time
Censoring Information
Uncensored value
Right censored value

Count
5
1

Censoring value: Censored = 0 = 0

Maximum Likelihood
Weibull

Estimation Method:
Distribution:

Parameter Estimates
Standard
Parameter Estimate
Error
Shape
8.91335
3.07276
Scale
989.515
50.6525
Log-Likelihood = -31.823

95.0% Normal CI
Lower
Upper
4.53525 17.5179
895.056 1093.94

Goodness-of-Fit
Anderson-Darling (adjusted) = 3.089
Characteristics of Distribution

Mean(MTTF)
Standard Deviation
Median
First Quartile(Q1)
Third Quartile(Q3)
Interquartile Range(IQR)

Estimate
936.625
125.579
949.652
860.433
1026.45
166.016

Standard
Error
52.8708
37.9430
52.9506
66.2041
51.5637
53.3979

95.0% Normal CI
Lower
Upper
838.527 1046.20
69.4595 227.040
851.341 1059.32
739.986 1000.49
930.202 1132.65
88.3823 311.841

Table of Percentiles

Percent
1
2
3
4
5
6
7
8
9
10
20
30
40
50
60
70
80
90

Percentile
590.588
638.711
668.820
691.158
709.092
724.172
737.242
748.818
759.237
768.733
836.256
881.437
917.684
949.652
979.858
1010.34
1043.78
1086.58

Standard
Error
114.953
107.747
102.734
98.8003
95.5254
92.7022
90.2104
87.9742
85.9418
84.0766
70.7763
62.4473
56.7372
52.9506
50.9185
50.7807
53.0748
59.5762

95.0% Normal CI
Lower
Upper
403.279 864.896
458.894 888.990
494.948 903.771
522.274 914.651
544.544 923.363
563.479 930.690
580.036 937.054
594.803 942.711
608.171 947.827
620.411 952.515
708.432 987.144
767.161 1012.74
812.955 1035.91
851.341 1059.32
884.973 1084.92
915.556 1114.93
944.772 1153.17
975.864 1209.85

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

STAT > REL. PARAMETRIC


RIGHT CENSORING
Distribution Analysis: Time

Weibull Rank Regression


Probability Plot for Time

Percent

Parameter Estimates
Table of Statistics
Shape
7.56295
Scale
988.765
Mean
928.587
StDev
145.190
Median
941.991
I QR
193.816
Failure
5
Censor
1
AD*
3.010
Correlation
0.971

90
80
70
60
50
40
30
20

Censoring Information Count


Uncensored value
5
Right censored value
1
Censoring value: Censored = 0 = 0
Estimation Method: Least Squares (failure time(X) on rank(Y))
Distribution:
Weibull

Weibull - 95% CI
Censoring Column in Censored = 0 - LSXY Estimates
99

Variable: Time

Parameter
Shape
Scale

Estimate
7.56295
988.765

Standard
Error
3.29313
59.5474

95.0% Normal CI
Lower
Upper
3.22146 17.7554
878.679 1112.64

Log-Likelihood = -31.929
Goodness-of-Fit
Anderson-Darling (adjusted) = 3.010
Correlation Coefficient = 0.971

10
Characteristics of Distribution

5
3
2
1

Time

Mean(MTTF)
Standard Deviation
Median
First Quartile(Q1)
Third Quartile(Q3)
Interquartile Range(IQR)

Estimate
928.587
145.190
941.991
838.588
1032.40
193.816

Standard
Error
61.4105
55.5404
62.0240
82.4504
63.1884
79.2151

95.0% Normal CI
Lower
Upper
815.698 1057.10
68.5997 307.293
827.943 1071.75
691.605 1016.81
915.697 1163.99
86.9949 431.804

Table of Percentiles

Percent
1
2

Percentile
538.188
590.240

Standard
Error
149.414
140.790

95.0% Normal CI
Lower
Upper
312.335 927.360
369.821 942.031

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

STAT > DISTRIBUTION ID

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

STAT > DISTRIBUTION ID


OUTPUT
Probability Plot for Time
LSXY Estimates-Censoring Column in Censored = 0
Weibull

Correlation Coefficient
Weibull
0.971
Lognormal
0.985
Exponential
*
Loglogistic
0.982

Lognormal
99
90

50

Percent

Percent

90

10

50
10

500

750
Time

1000

1250

600

Exponential

900
Time

1200

Loglogistic
99
90

50

Percent

Percent

90

10

50
10

10

100
Time

1000

10000

500

1000
Time

1500

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

STAT > DISTRIBUTION ID


SESSION OUTPUT

IE 7270: Reliability Estimation


Fall 2004

Distribution
Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

(adj) Coefficient
3.010
0.971
2.966
0.985
4.846
*
2.972
0.982
3.006
0.983
2.962
0.985
3.091
*
2.967
0.983
3.041
0.965
2.958
0.985
2.960
0.982

Table of Percentiles
Distribution
Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

Percent
1
1
1
1
1
1
1
1
1
1
1

Percentile
538.188
629.833
7.76524
589.331
749.089
611.877
771.599
569.976
437.715
579.636
518.562

Standard
Error
149.414
118.034
2.90518
126.776
5.91779
137.129
0.847374
145.153
239.707
172.859
196.889

Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

5
5
5
5
5
5
5
5
5
5
5

667.625
705.210
39.6310
693.414
757.883
697.362
779.260
686.298
634.819
683.597
668.119

124.906
99.1708
14.8270
102.899
19.9768
110.129
4.32469
111.843
162.917
129.462
134.911

462.681
535.325
19.0361
518.416
746.971
481.513
770.829
467.091
315.507
429.856
403.699

963.350
929.008
82.5071
927.485
798.066
913.212
787.782
905.506
954.131
937.339
932.539

Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

10
10
10
10
10
10
10
10
10
10
10

734.291
749.013
81.4052
746.392
769.478
745.393
789.303
743.418
721.865
739.019
735.819

108.985
87.9662
30.4558
89.8963
31.8354
95.3475
8.88327
95.4027
130.201
107.967
109.039

548.947
595.007
39.1017
589.450
746.971
558.515
772.082
556.432
466.676
527.407
522.105

982.213
942.881
169.476
945.119
834.474
932.270
806.907
930.404
977.054
950.631
949.532

Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

50
50
50
50
50
50
50
50
50
50
50

941.991
926.423
535.550
926.803
896.636
929.586
898.486
929.281
949.673
934.518
934.894

62.0240
64.8013
200.363
64.9904
75.0375
64.6516
58.4414
64.8025
60.3403
64.4412
64.4584

827.943
807.736
257.243
807.790
760.993
802.871
790.944
802.271
831.408
808.216
808.558

1071.75
1062.55
1114.95
1063.35
1056.46
1056.30
1020.65
1056.29
1067.94
1060.82
1061.23

Table of MTTF

Distribution
Weibull
Lognormal
Exponential
Loglogistic
3-Parameter Weibull
3-Parameter Lognormal
2-Parameter Exponential
3-Parameter Loglogistic
Smallest Extreme Value
Normal
Logistic

Mean
928.587
939.255
772.635
941.771
963.863
937.329
955.485
939.558
924.193
934.518
934.894

Standard
Error
61.411
67.278
289.063
67.159
116.291
65.727
84.313
65.914
65.442
64.441
64.458

95% Normal CI
Lower
Upper
312.335 927.360
436.220 909.379
3.72991 16.1663
386.588 898.398
746.971 760.778
343.109 880.644
769.939 773.261
285.481 854.472
-32.1025 907.532
240.840 918.433
132.666 904.457

95% Normal CI
Lower
Upper
815.698 1057.10
816.230 1080.82
371.123 1608.54
818.926 1083.04
760.882 1220.99
808.508 1066.15
803.735 1135.89
810.370 1068.75
795.929 1052.46
808.216 1060.82
808.558 1061.23

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

IE 7270: Reliability Estimation


Fall 2004

DISTRIBUTION OVERVIEW
WEIBULL RANK REGRESSION
Distribution Overview Plot for Time
LSXY Estimates-Censoring Column in Censored = 0
Probability Density Function

Table of Statistics
Shape
7.56295
Scale
988.765
Mean
928.587
StDev
145.190
Median
941.991
I QR
193.816
F ailure
5
Censor
1
AD*
3.010
Correlation
0.971

Weibull

0.003
90

PDF

Percent

0.002

0.001

0.000
500

750

1000

50

10

1
500

1250

750
Time

Time
Survival Function

1250

Hazard Function
0.03

100

0.02
Rate

Percent

1000

50

0
500

0.01

0.00
750

1000
Time

1250

500

750

1000
Time

1250

DEPARTMENT OF INDUSTRIAL &


MANUFACTURING ENGINEERING

You might also like