Indian Institute of Technology, Kharagpur

Department of Industrial and Systems Engineering

QDC Laboratory (IM39005)

Experiment No. 4: Experiment on Measurement of Process Capability

Group No. : 6

Group Members:

Name: Roll No.:

Aman Pramod Nagrale 19IM10004

Ananya Das 19IM10005
Nethala Revanth 19IM10020
Sanny 19IM10026

Date of Experiment/Data Collection: 6th September, 2021

Date of Submission: 9th September, 2021

Experiment Number: 4

Title of the Experiment: Experiment on Measurement of Process Capability

Objectives of the Study:

i) To understand the concept of process capability

ii) To measure capability of a process and draw inferences about the proces

Measuring Instruments/Apparatus/Software Required:

i) Vernier Caliper

ii) Micrometer

Experimental Procedures:
Process1. Manufacture ball bearings of different Diameters
1. Generate the ball diameters using random numbers using the distribution in the table given.
2. Consider USL = +3.5 σ , LSL = μ−3.5 σ, Target T =μ.

1
3. Calculate the required values. Tabulate the data as given below.
4. Plot the histograms for quality characteristics and infer regarding the process capability from
it relating to process target and variability.
capability = x ± 3 s
Process
5. Plot the normal probability plots (NPP) for quality characteristics and infer regarding the
process capability from it relating to process target and variability. Consider the mean of the
normal distribution as the 50th percentile for NPP plots; and the standard deviation of the
distribution as the slope of the NPP plots.

6. Calculate the indices for process capability,

C ,C ,C ,C ,C
p pu pl pk pm .
7. Finally, draw conclusions from the results obtained then discuss on them.

Process2. Manufacture a steel washer

1. Generate the dimensions using random numbers using the distribution given.
2. Consider
USL = +2.5 σ ,
LSL = μ−2.5 σ for inside diameter,
USL = + 4 σ , LSL = μ−4 σ for outside diameter,
USL = +2 σ , LSL = μ−2σ mm for thickness,
Target T =μ.
3. Calculate the required values. Tabulate the data as given below.
4. Plot the histograms for quality characteristics and infer regarding the process capability from
it relating to process target and variability.
capability = x ± 3 s
Process
5. Plot the normal probability plots (NPP) for quality characteristics and infer regarding the
process capability from it relating to process target and variability. Consider the mean of the
normal distribution as the 50th percentile for NPP plots; and the standard deviation of the
distribution as the slope of the NPP plots.

6. Calculate the indices for process capability,

C ,C ,C ,C ,C
p pu pl pk pm .
7. Finally, draw conclusions from the results obtained then discuss on them.

2
Data Collection and Analysis:

Sl. No. Process 1 Process 2

Ball Diameter Inside Diameter Outside Thickness

Diameter
1 2.1671 4.1103 6.9154 2.4178
2 03 2 95 77
3 2.4609 4.1806 7.2315 2.4256
4 3 16 44 5
2.6523 4.4453 7.3956 2.4264
5
35 21 37 94
6
3.0195 4.7209 7.7190 2.4430
7 24 64 23 92
8 3.0465 4.7210 7.8006 2.4453
9 12 09 71 5
10 3.0479 4.7611 7.8105 2.4498
11 66 61 43 95
12 3.1126 4.8930 7.8416 2.4562
13 53 06 61 98
14 3.1289 4.9211 7.8505 2.4582
15 03 71 38 71
16 3.1878 5.0307 7.9338 2.4614
21 7 46 84
17
3.1926 5.0532 8.0650 2.4654
18 12 05 69 68
19 3.1983 5.0811 8.2062 2.4667
20 99 49 41 62
21 3.2106 5.0853 8.2275 2.4683
22 45 99 31 2
23 3.2123 5.1669 8.2329 2.4696
24 17 51 04 75
25 3.2179 5.1910 8.2552 2.4716
26 14 44 84 69
27 3.2289 5.2276 8.2958 2.4730
78 85 37 69
28
3.2508 5.3078 8.3104 2.4761
29
14 08 77 3
30 3.3084 5.3473 8.3177 2.4787
31 45 99 03 43
32 3.3496 5.3843 8.3219 2.4801
33 53 9 19 82
34 3.3765 5.3873 8.3221 2.4812
35 23 1 21 44
36 3.4019 5.4008 8.3247 2.4917
37 68 02 91 24

3
38 3.4129 5.4079 8.3248 2.4922
39 31 99 45 44
40 3.4715 5.4301 8.3562 2.4963
41 87 07 26 67
3.5234 5.4641 8.3570 2.4970
42
86 98 45 02
43
3.5266 5.4880 8.4691 2.5016
44 62 48 44 88
45 3.5529 5.4894 8.4705 2.5044
46 01 17 43 4
47 3.5593 5.5052 8.4919 2.5049
48 8 84 14 39
49 3.5930 5.5060 8.5195 2.5071
50 12 97 84 78
3.6012 5.5243 8.5285 2.5096
4 57 9 92
3.6027 5.5283 8.5965 2.5102
13 88 99 02
3.6198 5.6933 8.6385 2.5106
61 81 71 16
3.6209 5.7121 8.6386 2.5132
23 37 66 02
3.6231 5.7163 8.6508 2.5152
62 35 94 24
3.6539 5.7352 8.6708 2.5193
77 14 38 86
3.6850 5.7697 8.6847 2.5196
95 55 8 24
3.7060 5.7721 8.6866 2.5243
53 8 9 54
3.7379 5.7794 8.7032 2.5251
91 96 96 2.5256
3.7422 5.7942 8.7273 46
91 77 44 2.5302
3.8302 5.8242 8.7423 2
23 74 44 2.5312
3.8411 5.8571 8.7610 01
84 63 29 2.5374
3.8503 5.8761 8.7722 01
91 05 74 2.5374
3.9340 5.8956 8.7752 41
29 65 06 2.5375
3.9455 5.9396 8.8631 63
38 53 51 2.5401
3.9544 5.9728 8.9041 7
92 61 81 2.5420

4
 4.0829 5.9894 8.9146 5
28 94 59 2.5498
4.1227 6.0450 9.0709 31
51 29 53 2.5558
4.1442 6.1717 9.2931 89
27 69 19 2.5580
4.2150 6.4912 9.3724 9
51 03 83 2.5612
4.2428 6.6241 9.3999 31
04 51 95 2.5707
4.2502 6.6739 9.5164 67
03 91 18 2.5876
4.4195 6.7242 9.8724 31
24 35 06

Calculation of Natural Tolerance Limits ( X́ ± 3 s):

Ball Diameter ( X́ ¿: 3.516773 UNTL= 4.865043 LNTL = 2.168502

Washer Inside Diameter ( X́ ¿: 5.496395 UNTL= 7.191422 LNTL = 3.801367

Washer Outside Diameter ( X́ ¿: 8.463052 UNTL= 10.11878 LNTL = 6.807321

Washer Thickness ( X́ ¿: 2.500476 UNTL= 2.618627 LNTL = 2.382324

1. Histograms for quality characteristics (With UNTL, LNTL, USL, LSL)

1.1 Process 1: Histogram for Ball Diameter

5
1.2 Process 2: Histogram for Inside Diameter

1.2 Process 2: Histogram for Outside Diameter

6
1.4 Process 2: Histogram for Thickness

2. Normal probability plots (NPP) for quality characteristics

2.1 Process 1: Normal probability plot for Ball Diameter

7
2.2 Process 2: Normal probability plot for Inside Diameter

2.3 Process 2: Normal probability plot for Outside Diameter

8
2.4 Process 2: Normal probability plot for Thickness

3. Process capability indices calculation (for all process variables)

C =
3.1 p Ball = 1.297959 Outer Diameter = 1.207926
Inner Diameter = 0.737451 Thickness = 0.84637

3.2 Cpu = Ball = 1.21135 Outer Diameter = 1.230241

Inner Diameter = 0.739578 Thickness = 0.842344

9
3.3 Cpl = Ball = 1.384568 Outer Diameter = 1.185611

Inner Diameter = 0.735324 Thickness = 0.850396

C
3.4 pk = Ball = 1.21135 Outer Diameter = 1.185611
Inner Diameter = 0.735324 Thickness = 0.842344

3.5
C pm = Ball = 1.256246 Outer Diameter = 1.205228
Inner Diameter = 0.737436 Thickness = 0.846308

4. Results and Discussions:

a) What is the inference on process capability based on histograms for process 1

and 2?

For Process 1, the Lower Natural Tolerance Limits (LNTL) and Upper Natural Tolerance Limits
(UNTL) are within the Upper and Lower Specification range i.e., LSL and USL. The control
limits, which are a measure of variability of the sample means are well within the Specification
limits (acceptable bounds of variability for individual items). Hence, Process 1 is capable in
terms of its operations.

For Process 2, the Lower Natural Tolerance Limits (LNTL) and Upper Natural Tolerance Limits
(UNTL) are outside the Upper and Lower Specification range i.e., LSL and USL (process
spread). Hence, Process 2 is not capable in terms of its manufacturing of the Inner Diameter and
Thickness.

Whereas, for Outside diameter, the LNTL and UNTL and within the range of process spread
(within bounds of USL and LSL). Hence, the manufacturing of Outside diameter is a capable
process and is within acceptable bounds of variability.

b) What is the inference on process capability based on NPPs for process 1 and 2?

NPP for Process-1 looks fairly straight, at least when the few large and small values are ignored.
Hence, the process is capable as it follows normal distribution. We observe a few values towards
the extreme, but they can be ignored.

10
NPP for process-2 is also fairly straight for Inside Diameter, Outside Diameter and Thickness.
Hence process-2 observations are normally distributed.
Hence, according to NPP, both process 1 and process2 are capable.

c) What is the inference on process capability based on the process capability

indices for process 1 and 2?

For Process-1 (Ball diameter), Cp = 1.297959 > 1: Thus, process can potentially meet
specifications; Cpu = 1.21135, Cpl=1.384568, this means Cpl and Cpu are >1 hence the process is
very capable and a negligible proportion of non- confirming products are there beyond the
ranges. Cpk=1.21135 >1. Hence, the process mean is centered at the target value. Cpm= 1.256246
>1 Hence variability of the process around the Target value is less.

So, Process-1 is capable.

For Process-2 (ID), Cp=0.737451 < 1: Thus, process can potentially meet specifications;
Cpu=0.739578, Cpl=0.735324, this means Cpl and Cpu are < 1, which is not desirable, hence the
process is not very capable and a significant proportion of non-confirming products are there
beyond the ranges. Cpk=0.735324 <1 Hence, the process mean is not centered at the target value.
Cpm= 0.737436 <1 Hence variability of the process around the Target value is larger than
acceptable. So, Process-2 is incapable.
Since the process of carving the ID is incapable hence, the process for thickness is also incapable
(as visible from the Process Capability Indices).
However, the process capability indices for Process 2 (OD) are all greater than 1. Hence the
process can potentially meet specifications and is also centered at the target. So, process 2 (OD)
is capable.

Overall Process 2 is incapable.

d) Compare the inferences on process capability based on histograms, NPPs and

process capability indices for process 1 and 2

Process 1:
According to Histograms, NPP and Process Capability Indices, Process 1 is Capable. So, we can
conclude that Process 1 is indeed capable.

Process 2:

11
According to Histograms and Process Capability Indices Process for ID and T are incapable,
whereas according to NPP Process for ID and T are normally distributed and process is capable.
Hence, NPP is not that reliable in terms of telling us if a process is capable or not.

Process for OD is capable according to Histograms, NPP and Process Capability Indices.

5. Conclusion:

By this experiment we are come to know about the main indicators that monitor processes
(Process Capability, Process Performance and Process Sigma) aimed at determining the defect
rates of a process, and, therefore, at quantifying defective parts per million (ppm). This is aimed
at any organizations wanting to achieve high standards of quality and process yield. Indeed, by
keeping process under statistical control, it is possible to have more robust information on the
process, with more reliable indicators, which also allows to better satisfy customers’
requirements. Process Capability indices can be used to analyze existing processes, and only
when the analyzed process in under statistical control. Process Performance indices are used to
analyze new processes, or when the processes are not under statistical control. Process Sigma,
instead, is likely to create one common metric that can be applicable to all data and
environments. This discussion puts forward another crucial implicit assumption that is data
distribution for all above- mentioned indices needs to be Normal. If the distribution happens not
to be Normal, these indices can be calculated, but the information obtained is not reliable.

12