Dr.

Akhilesh Das Gupta Institute of

Technology and Management

STATICAL QUALITY CONTROL AND RELIABILITY

Practical File

Submitted By :-
ASHISH KUMAR
35115608218
Mechanical Engineering (4th Year)

Submitted To :-
Mr. Rajat Gupta
Question 1 - A semiautomatic turret lathe machines the thickness of a pan that is
subsequently used in an assembly. The process mean is known to be 30 mm with a
standard.
Average Part Thickness Values

Average Part Average Part Average Part

Sample Thickness, * (mm) Sample Thickness, * (mm) Sample Thickness, * (mm)
1 31.56 6 31.45 11 30.20
2 29.50 7 29.70 12 29.10
3 30.50 8 31.48 13 30.85
4 30.72 9 29.52 14 31.55
5 28.92 10 28.30 15 29.43

deviation of 1.5 mm. Construct a control chart for the average thickness using 3o limits if
samples of size 5 are randomly selected from the process. Table 6-1 shows the average
thickness of 15 samples selected from the process. Plot these on a control chart, and
make inferences.

Solution 1 - The centre line is

CL = 30 mm
The standard deviation of the sample mean I is given by

= 0.671 mm
Assuming a normal distribution of the sample mean thickness, the value of k is selected as
3. The control limits are calculated as follows:

UCL — 30 + (3)(0.671) = 32.013

LCL — 30 — (3)(0.671) = 27.987

The centerline and control limits are shown in Figure. The sample means for the 15 samples
shown in Table are plotted on this control chart. Figure shows that all of the sample means
are within the control limits. Also, the pattern of the plot does not exhibit any nonrandom
behavior. Thus, we conclude that the process is in control.
Question 2 - The thickness of the magnetic coating on audio tapes is an important
characteristic. Random samples of size 4 are selected, and the thickness is measured using
an optical instrument. Table shows the mean x̄ and standard deviation s for 20 samples.
The specifications are 38 ± 4.5 micrometers (μm). If a coating thickness is less than the
specifications call for, that tape can be used for a different purpose by running it
through another coating operation.
Sample Mean, Sample Standard Sample Mean, Sample
Sample x̄ Deviation, s Sample x̄ Standard
Deviation, s
I 36.4 4.6 11 36.7 5.3
2 35.8 3.7 12 35.2 3.5
3 37.3 5.2 13 38.8 4.7
4 33.9 4.3 14 39.0 5.6
5 37.8 4.4 15 35.5 5.0
6 36.1 3.9 16 37.1 4.1
7 38.6 5.0 17 38.3 5.6
8 39.4 6.1 18 39.2 4.8
9 34.4 4.1 19 36.8 4.7
10 39.5 5.8 20 37.7 5.4

(a) Find the trial control limits for an x̄ - and an s-chart.

Solution- The standard deviation chan must first be constructed. The centerline of the
s-chart is
CLi = s = 95.80/20 = 4.790
The control limits for the s-chart are
UCLs = B4s = (2.266) (4.790) = 10.854
LCLs = B3s = (0) (4.790) = 0
Figure shows this standard deviation control chart. None of the points fall outside the
control limits, and the process seems to be in a state of control, so the k-chart is
constructed next. The centerline of the x̄-chart is
CLx̄ = x̄ = 743.5/20 = 37.175

The control limits for the x̄-chart are

UCLx̄ = x̄ + A3s = 37.175 + (1.628)(4.790) = 44.973 LCLx̄ =
x̄ — A3s = 37.175 — (1.628)(4.790) = 29.377

depicts the x̄-chart. All the points are within the control limits, and no unusual nonrandom
pattern is visible on the plot.

(a) Assuming special causes for the out-of-control points, determine the revised control
limits.

Solution- In this case, the revised control limits will be the same as the trial control limits
because we believe that no special causes are present in the system.

(b) Assuming the thickness of the coating to be normally distributed, what proportion of the
product will not meet specifications?

Solution- The process standard deviation may be estimated as

σ = s/c4 = 4.790/ 0.9213 = 5.199
To find the proportion of the output that does not meet specifications, the standard normal
values at the upper and lower specification limits (USL and LSL) must be found. At the lower
specification limit we get
z1 = (33.5-37.175)/5.199 = -0.71

The area below the LSL, found by using the standard normal table in Appendix A-3, is
0.2389. Similarly, the standard normal value at the upper specification limit is
z2 = (42.5-37.175)/5.199 = 1.02
From Appendix A-3, the area above the USL is 0.1539. Hence, the proportion of product
not meeting specifications is 0.2389 + 0.1539 = 0.3928.
 (d) Comment on the ability of the process to produce items that meet specifications.

Solution- A proportion of 39.28'7o of product not meeting specifications is quite high. On the
other hand, we found the process to be in control. This example teaches an important lesson. It
is possible fora process to be in control and still not produce conforming items. In such
cases, management must look for the prevailing common causes and come up with ideas for
process improvement. The existing process is not capable of meeting the stated
specifications.

(e)If the process average shifts to 37.8 pm, what proportion of the product will be
acceptable?

Solution- If the process average shifts to 37.8{im, the standard normal values must be
recalculated. At the LSL,
z1 = (33.5-37.8)/5.199 = -0.83
From the standard normal table in Appendix A-3, the area below the LSL is 0.2033. The
standard normal value at the USL is
z2 = (42.5-37.8)/5.199 = 0.90
The area above the USL is 0.1841. So, the proportion nonconforming is 0.2033 +
0.1841 = 0.3874. Although this change in the process average does reduce the proportion
nonconforming, 38.74% nonconforming is still quite significant.

Question 3 - Management has decided to set a standard of 3% for the proportion of

nonconforming test tubes produced in a plant. Data collected from 20 samples of size 100
are shown in Table, as is the proportion of nonconforming test tubes for each sample. The
centerline and control limits, based on the specified standard, are found to be
CLp = p0 = 0.030
UCLp = 0.030 + 3√(0.03)(1-0.03)/100 = 0.081
LCLp = 0.030-3√(0.03)(1-0.03)/100 = -0.021→0
A p-chart using these limits is shown in Figure. From the figure we can see that the process is
out of control; samples 8 and 11 plot above the upper control limit. The special cause for sample
8 is improper alignment of the die, for which remedial action is taken. For sample 11, no special
cause is identified. Hence, the process is viewed to be out of control with respect to the current
standard. From Figure, observe that the general level of the proportion nonconforming does not
match the standard value of 3%; the actual proportion nonconforming appears higher. Using the
data from Table, the process average proportion nonconforming is
P = 84/2000 = 0.042

Number of Number of
Items Nonconforming Proportion
Sample Date Time Inspected, n Items, z Nonconforming, p Comments
1 9/8 8:20 100 4 0.04
2 9/8 8:45 100 2 0.02
3 9/8 9:10 100 5 0.05
4 9/8 9:30 100 3 0.03
5 9/9 9:00 100 6 0.06
6 9/9 9:20 100 4 0.04
7 9/9 9:50 100 3 0.03
8 9/9 10:20 100 9 0.09 Die not
9 9/10 9:10 100 5 0.05 aligned
10 9/10 9:40 100 6 0.06
11 9/10 10:20 100 9 0.09
12 9/10 10:45 100 3 0.03
13 9/11 8:30 100 3 0.03
14 9/11 8:50 100 4 0.04
15 9/11 9:40 100 2 0.02
16 9/11 10:30 100 5 0.05
17 9/12 8:40 100 3 0.03
18 9/12 9:30 100 I 0.01
19 9/12 9:50 100 4 0.04
20 9/12 10:40 100 3 0.03
2000 84

This value exceeds the desired standard of 3%. From Figure only three points are below the
standard of 3%. This confirms our suspicion that the process mean is greater than the desired
standard value. If sample 8 is eliminated following removal of its special cause, the revised
process average is
P = (84-9)/1900 = 0.039
When the control limits are calculated based on this revised average, we have
CLp = 0.039
UCLp = 0.039 + 3√(0.039)(1-0.039)/100 = 0.097
LCLp = 0.039 - 3√(0.039)(1-0.039)/100 = -0.019→0

If a control chart were constructed using these values, the remaining samples (including sample
11) would indicate a process in control, with the points hovering around 0.039, the calculated
average.
With the standard of 3% in mind, we must conclude that the process is currently out of control
because sample 11 would plot above the upper control limit value of 0.081, as found using the
standard. We have no indication of the special causes for the out-of-control point (sample 11), so
we cannot take remedial action to bring the process to control. Furthermore, the process average
proportion nonconforming should be reduced; it is too far from the desired value of 3%.
Actions to accomplish this task must originate from management and may require major changes
in the product, process, or incoming material quality. Operator-assisted measures are not
sufficient to bring the process to control.

Question 4 - The diameter of a part has to fit an assembly. The specifications for the
diameter are 5 ± 0.015 cm. The samples taken from the process in control yield a sample mean
x̄ of 4.99 cm and a sample standard deviation s of 0.004 cm. Find the natural tolerance limits of
the process. Would you consider adjusting the process centre?
Solution - The upper and lower natural tolerance limits based on the sample estimates are found
using eq.:
UNTL = 4.99 + (3)(0.004) = 5.002
LNTL = 4.99 - (3)(0.004) = 4.978
Assuming a normal distribution of diameters, the process spread is (6)(0.004) = 0.024 cm, which is the
difference between the natural tolerance limits. For the current process, we would expect the
diameters to lie between 5.002 and 4.978 cm.
The difference between the specification limits is 0.03 cm. If the process were left in its original
state, some proportion of the parts would fall below the lower specification limit of 4.985 cm. To
calculate this proportion, the standardized normal value at the LSL must first be found as follows:
Z = (LSL- x̄)/s = (4.985 - 4.99)/0.004 = -1.25
This Z-value is found in Appendix A-3, the proportion below the LSL is 0.1056. Thus, it would be
desirable to adjust the process center to the target value of 5 cm. If this is done, since the process
spread is 0.024 cm and the difference between the specification limits is 0.03 cm, virtually all parts
would fall between the specification limits, and we would have a capable process.
Statistical tolerance limits are the limits of an interval that (with a given level of confidence ϒ)
contains at least a specified proportion (1 - α) of the population. These limits are found from sampling
information. For example, if we conclude, using a level of confidence of 0.98 and samples of size 10,
that 95% of the part lengths fall between 30 and 35 mm, the statistical tolerance limits are 30 and 35
mm. Sample estimates are used to infer population parameters; the limits are influenced by the
sample size. As the sample size becomes large, the statistical tolerance limits approach the values
found using the population parameters. Estimation of statistical tolerance limits is discussed in
Section 9-10. Estimation of these limits is based on a normal distribution and on nonparametric
methods.

Question 5 - Construct an OC curve for a single sampling plan where the lot size is 2000, the sample
size is 50, and the acceptance number is 2.
Solution - We are given N = 2000, n = 50, and c = 2. The probability of lot acceptance is equivalent to
the probability of obtaining 2 or fewer nonconforming items in the sample. The Poisson probability
distribution in Appendix A-2 is used to obtain the lot acceptance probability for different values of
the proportion nonconforming p. Let’s suppose that p is 0.02 (i.e., the batch is 2% nonconforming).
Since np = (50) (0.02) = 1.0, the probability Pa of accepting the lot (using Appendix A-2) is 0.920. Table
shows values of a for various values of p. In some instances, the probability values are linearly
interpolated from Appendix A-2. A plot of these values, the OC curve, is shown in Figure.
The discriminating power of the sampling plan N = 2000, n = 50, c = 2 can be seen from the OC curve
in Figure. If a series of batches, each of which is l% nonconforming, comes in for inspection,
then (using this plan) the probability of lot acceptance is 0.986. It means that, on average, about 986
out of 1000 such batches will be accepted by the sampling plan. On the other hand, if batches are
5% nonconforming, only about 544 out of 1000 batches will be accepted. As the lot quality
becomes poorer, the probability of lot acceptance decreases, as it should. The steeper the drop in
the probability of lot acceptance as lot quality worsens, the higher the discriminatory power
of the sampling plan.
Producer and consumer risk can also be demonstrated through the OC curve. Suppose that our
numerical definition of good quality (indicated by the AQL) is 0.01 and that of poor quality
(indicated by the LQL) is 0.11. From the OC curve in Figure, the producer’s risk α is 1 - 0.986 = 0.014.
We consider batches that are 1% nonconforming to be good. If our sampling plan is used, such batches
will be rejected about 1.4% of the time. Batches that are 11% nonconforming, on the other hand, will
be accepted 8.8% of the time. The consumer’s risk is therefore 8.8%.

Effect of the Sample Size and the Acceptance Number

The parameters n and c of the sampling plan affect the shape of the OC curve. As long as the lot size N
is significantly large compared to the sample size n, the lot size does not have an appreciable impact
on the shape of OC curve. For fixed values of N and c, as the sample size becomes larger, the slope of
the OC curve becomes steeper, implying a greater discriminatory power.
Figure shows the OC curves for those sampling plans. Note that, for lots that are 2% nonconforming,
the sampling plan N = 2000, n = 50, c = 2 will accept such lots about 92% of the time. However, for the
same lots, the sampling plan N = 2000, n = 200, c = 2 will accept them only 23.8% of the time. Changing
the sample size from 50 to 200 causes a drop in the acceptance probability of 68.2%.
For fixed values of the lot size N and the sample size n, as the acceptance number
decreases, the slope of the OC curve becomes steeper. Figure shows the OC curves for four sampling
plans. Note that the probability of acceptance decreases for a given lot quality as the acceptance
number c decreases.

0.1
6

One comment needs to be made regarding the case where the acceptance number is zero. The OC
curve starts dropping drastically even as the proportion nonconforming deviates slightly from
zero. This may not be desirable from a producer’s point of view. For example, if lots that are 0.5%
nonconforming are considered acceptable, the sampling plan N = 2000, n = 50, c = 0 will reject such
lots about 22% of the time, implying a high value of the producer’s risk. Sampling plans with c = 0 do
not have the desirable inverted-S shape of the ideal OC curve. They are, however, stringent and
serve a need.
The chosen values of n and c should be such that they match the goals of the user. Given some
desirable producer’s risk and the associated quality level of a good lot (AQL) and/or a desirable
consumer’s risk and an associated quality level of a poor lot (LQL), the combination of n and c
that produces an OC curve that matches these goals will provide an acceptable sampling plan.