JIMMA UNIVERSITY

Jimma Institute of Technology

Post Graduate Study

Faculty of Mechanical Engineering

MSc. In Mechanical System Design Engineering

Product Testing Equipment

Assignment-4

By:

1) Tesfaye Olana RM0978/09

Submitted to: Chalachew S. (MSc)

Jan, 2018
 PTE-Assignment: 3

Exercise on chapter-7

1. A factory produces batches of bars (each batch contains 1000 bars). For each batch 100

samples are taken to evaluate the diameter of the bars according to the customer’s

requirement given below. Three batches are tested. The test results are indicated in table

1.

Table 1: The critical dimension requiring inspection (diameter of the bar) is 200.02 mm.

Acceptable size: 19.98-20.02 mm.

First Batch Second Batch Third Batch

Diameter (mm)
Number of Bars Number of Bars Number of Bars
19.96 1 0 1
19.97 3 2 1
19.98 14 10 15
19.99 18 22 16
20.00 20 30 31
20.01 22 25 20
20.02 12 10 12
20.03 10 1 4

For the data given above,

a) Calculate the acceptable quality level (AQL) for each batch of samples.
b) Identify the unacceptable batches.
c) Suggest the next correction steps to be taken for the unacceptable batches.

Solution:

To determine the AQL of each batch first let determine each batch’s defects based on the

given information:

First Batch
Diameter Second Batch Third Batch
Number of
(mm) Number of Bars Number of Bars
Bars
19.96 1 0 1
19.97 3 2 1

By: Tesfaye Olana, 2018 1

 PTE-Assignment: 3

19.98 14 10 15
19.99 18 22 16
20.00 20 30 31
20.01 22 25 20
20.02 12 10 12
20.03 10 1 4
Total Number of defects 1+3+10=14 0+2+1=3 1+1+4=6

a) Calculate the acceptable quality level (AQL) for each batch of samples.

i) For the first batch of bar;

𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡 𝑖𝑛 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ

Acceptable quality level (AQL) = 𝑡𝑜𝑡𝑎𝑙 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ
𝑥100

14
= 𝑥100 = 14%
100

ii) For the second batch of bar;

𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡 𝑖𝑛 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ

Acceptable quality level (AQL) = 𝑡𝑜𝑡𝑎𝑙 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ
𝑥100

3
= 𝑥100 = 3%
100

iii) For the third batch of bar;

𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑑𝑒𝑓𝑒𝑐𝑡 𝑖𝑛 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ

Acceptable quality level (AQL) = 𝑥100
𝑡𝑜𝑡𝑎𝑙 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑎𝑡𝑐ℎ

6
= 100 𝑥100 = 6% ,

b) Identify the unacceptable batches.

So from the result determined in solution (a) for the three batches the acceptance quality

level is given by for all batch 1, 2 & 3, is equal to 14%, 3% and 6% respectively.

Since the acceptance figure is given by AQL<5%, the first and third batch is failed and only

the second batch will satisfy acceptance figure. So the second batch is acceptable.

By: Tesfaye Olana, 2018 2

 PTE-Assignment: 3

c) Suggest the next correction steps to be taken for the unacceptable batches.

The next correction steps for unacceptable batches; inspect the second sample of the first

and third batches of bars to confirm the findings of the first sample of first and third

batches of bars. If the second batches taken also contains more than 5% defective

components then: Carry out 100% inspection of the first and third batches of bars to

remove all rejects and rectify them if possible which is highly costly. Sell the product, the

first and third bars to a customer who will accept the higher AQL, possibly at a lower price

or by discount.

Question 2. An NGO is interested in evaluating the consequences of a 20,000hr guarantee

on newly acquired water pumps. 16 of the pumps are selected for purposes of registration.

As shown in table 2, at a certain point in time 10 pumps failed (pumps 2, 4, 5,7,8,9, 11,12,13

and 15), while the rest had been withdrawn from the registration scheme for emergency use

in drought-affected areas at the indicated times. The data is as shown in table 2.

Pump No. Pump Test Time (hr.) Pump No. Pump Test Time (hr.)

1 10,000 9 16,800 (Failed)

2 10,100 (Failed) 10 17,000

3 11,000 11 18,000 (Failed)

4 12,200 (Failed) 12 18,500 (Failed)

5 13,500 (Failed) 13 19,000 (Failed)

6 14,600 14 20,000

7 15,700 (Failed) 15 21,000 (Failed)

8 16,000 (Failed) 16 21,500

By: Tesfaye Olana, 2018 3

 PTE-Assignment: 3

Using the above test data in table 2

a. Fill the incomplete table given below (table3), i.e. calculate rank order, median rank and

95% rank for failed pumps.

b. Determine the Shape Factor and Characteristic Life for the given data.

c. Plot a percentage failure verses age-at-failure line (Weibull plot).

d. Determine the expected time when 20% of the population (pumps) will have failed.

e. What percentage of the population (pumps) may have failed at the guaranteed time?

Solution:

a. Fill the incomplete table given below (table3), i.e. calculate rank order, median

rank and 95% rank for failed pumps.

𝑖−0.3
Median rank= 𝑛+0.4 𝑤ℎ𝑒𝑟𝑒 𝑖 𝑖𝑠 𝑟𝑎𝑛𝑘 𝑜𝑟𝑑𝑒𝑟, 𝑛 𝑖𝑠 10 𝑖𝑛 𝑡ℎ𝑒 𝑔𝑖𝑣𝑒𝑛 𝑠𝑎𝑚𝑝𝑙𝑒 𝑎𝑛𝑑

𝑖−0.3
𝑡ℎ𝑒 𝑟𝑎𝑛𝑘(%) = 𝑥100. The calculated value of item is tabulated as below in table 3.
𝑛+0.4

Failure Time Rank Order Median Rank (%) 95% Rank (%)
10100 1 6.730769231 6.394230769
12200 2 16.34615385 15.52884615
13500 3 25.96153846 24.66346154
15700 4 35.57692308 33.79807692
16000 5 45.19230769 42.93269231
16800 6 54.80769231 52.06730769
18000 7 64.42307692 61.20192308
18500 8 74.03846154 70.33653846
19000 9 83.65384615 79.47115385
21000 10 93.26923077 88.60576923

b. Determine the Shape Factor and Characteristic Life for the given data.

To determine the shape factor and characteristics life we have to work the following

analysis on Excel

By: Tesfaye Olana, 2018 4

 PTE-Assignment: 3

For t=0, failure will become

𝑡
−( )𝛽
𝐹(𝑡) = 1 − 𝑒 𝜂

𝑡 𝛽
1 ( )
LET 𝐷 = 1−𝐹(𝑡) = 𝑒 𝜂 , 𝐸 = ln(ln(𝐷)) = 𝛽 ln(𝑡) − 𝛽 ln(η)

𝑖𝑓 ln(𝑡) = 𝑥, 𝐸 = 𝑦, 𝑎 = −𝛽 ln(η), 𝑏 = 𝛽

𝑡ℎ𝑒𝑛 𝑤𝑒 𝑤𝑖𝑙𝑙 ℎ𝑎𝑣𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑡𝑟𝑎𝑖𝑔ℎ𝑡 𝑙𝑖𝑛𝑒 𝑓𝑖𝑡 𝑡𝑜 𝑦 = 𝑏𝑥 + 𝑎

Assuming F(t)= median , t= design cycles , calculate D, E and x

failure time rank order median 95% rank (%) E=ln(ln(D)) X=ln(t)

10100 1 0.067307692 1.018556701 -3.99613 9.220291

12200 2 0.163461538 1.135632184 -2.06208 9.409191
13500 3 0.259615385 1.283116883 -1.38913 9.510445
15700 4 0.355769231 1.474626866 -0.94571 9.661416
16000 5 0.451923077 1.733333333 -0.59775 9.680344
16800 6 0.548076923 2.10212766 -0.29713 9.729134
18000 7 0.644230769 2.67027027 -0.01798 9.798127
18500 8 0.740384615 3.659259259 0.260255 9.825526
19000 9 0.836538462 5.811764706 0.565248 9.852194
21000 10 0.932692308 14.11428571 0.973498 9.952278

Using the data analysis tool bar on excel perform the analysis and excel will give the

following output,

SUMMARY X Variable 1 Line Fit Plot

OUTPUT 2

0
Regression Statistics
9 9.2 9.4 9.6 9.8 10
Multiple R 0.987849181 -2 Y
Y

R Square 0.975846004 -4 Predicted Y

Adjusted R Square 0.972826754
Standard Error 0.240710829 -6
X Variable 1
Observations 10

By: Tesfaye Olana, 2018 5

 PTE-Assignment: 3

ANOVA
Significa
df SS MS F nce F
Regres 18.7272 18.72722 323.208 9.3984E-
sion 1 2963 963 1308 08
Residu 0.46353 0.057941
al 8 3627 703
19.1907
Total 9 6325
Coefficien Standard Lower Upper Lower Upper
ts Error t Stat P-value 95% 95% 95.0% 95.0%
- - - - - -
Interce 62.92034 3.45893 18.19066 8.57181 70.8966 54.9440 70.8966 54.9440
pt 022 5548 57 E-08 599 2054 599 2054
X
Variabl 6.433187 0.35783 17.97799 9.3984E- 5.60801 7.25836 5.60801 7.25836
e1 886 6878 018 08 4565 1208 4565 1208
RESIDUAL
OUTPUT

Observation Predicted Y Residuals

1 -3.604477762 -0.391653951
2 -2.389245174 0.327167886
3 -1.737860881 0.348731236
4 -0.766635899 -0.179070808
5 -0.644868455 0.0471157
6 -0.330992162 0.033865635
7 0.112851943 -0.130832944
8 0.289114692 -0.028859774
9 0.460676536 0.10457151
10 1.104532229 -0.13103449

From the above tables the intercepts

a=-62.92034022 and b=6.433187886=β

62.92034022
𝜂 = 𝑒 −𝑎/𝛽 = 𝑒 −(−6.433187886) = 17686.99473

𝜂 = 17687

By: Tesfaye Olana, 2018 6

 PTE-Assignment: 3

c. Plot a percentage failure verses age-at-failure line (Weibull plot).

Using the above shape parameters and characteristic lives, the reliability or failure can be

calculated by using the previous equations. We can also use weibull function in Microsoft

Excel. Weibull(cycles, beta, etha, true) =Failure probability

Beha 17687
Etha 6.433187886
Cycles Failures Probability Reliability
2000 8.13517E-07 0.999999186
4000 7.02874E-05 0.999929713
6000 0.000953851 0.999046149
8000 0.006054809 0.993945191
10000 0.025194827 0.974805173
12000 0.079146538 0.920853462
14000 0.199301465 0.800698535
16000 0.408291072 0.591708928
18000 0.673544078 0.326455922
20000 0.889725994 0.110274006
22000 0.982930833 0.017069167
24000 0.999194669 0.000805331
26000 0.99999336 6.63989E-06
28000 0.999999995 4.56552E-09
30000 1 9.99815E-14
Plotting this on excel tool

Weibull Distribution
1.2

0.8

0.6

0.4

0.2

0
0 5000 10000 15000 20000 25000 30000 35000
-0.2

failures probability reliability

By: Tesfaye Olana, 2018 7

 PTE-Assignment: 3

d. Determine the expected time when 20% of the population (pumps) will have failed.

𝑡 𝛽
−( )
𝑅(𝑡) = 𝑒 𝜂
1
1 1 1 6.433187886
𝑡 = η(ln( ))𝛽 = 17687(ln ( ) = 19045
𝑅(𝑡) 0.2

e. What percentage of the population (pumps) may have failed at the guaranteed

time?

𝑡 𝛽
−( )
𝑅(𝑡) = 𝑒 𝜂

20000 6.433187886
−( )
𝑅(20000) = 𝑒 17687 = 11.026%

……………………………………………..

The End!

By: Tesfaye Olana, 2018 8