STATISTICAL

QUALITY CONTROL
For Variables
Courtesy: Dr. Pravin Kumar
Introduction
• Control charting is one of the tools of Statistical Quality Control (SQC).
• It is the most technically sophisticated tool of SQC. Dr. Walter A. Shewhart (1920)
of the Bell Telephone Labs developed it as a statistical tool to the study of
manufacturing process variation for the purpose of improving the economic
effectiveness of the process.
• These methods are based on continuous monitoring of process variation.

Industrial Engg. & Management, 1/e, 2

Pravin kumar, Pearson Education
Industrial Engg. & Management, 1/e, 3
Pravin kumar, Pearson Education
Sources of Variations
• Variation due to process includes tool wear, way of operation, tool vibration, jigs
and fixtures, etc.
• Variation due to materials includes quality characteristics of raw materials, such
as tensile strength, ductility, malleability, hardness, toughness, thickness, porosity
and moisture contents.
• Variation due to environment includes temperature, light, radiation, humidity etc.
Variation due to operator includes skill, motivation, wages etc.
• Finally, variation due to inspection includes incorrect application of quality
standards and inspection device.

Industrial Engg. & Management, 1/e, 4

Pravin kumar, Pearson Education
Types of Variations
• Within piece variation: This type of variation in quality parameters varies with in
piece, for example, surface roughness of the different faces of the piece.
• Piece-to-Piece Variation: This type of variation in quality parameters varies from
piece to piece, for example, change in surface finish due to continuous tool wear.
• Time-to-time variation: This type of variation in quality parameters varies due to
different time of the day, for example, product produced in the morning will be
better than the product produced after launch due to laziness of the worker.

Industrial Engg. & Management, 1/e, 5

Pravin kumar, Pearson Education
 Chart Techniques
• Select the quality characteristics
• Choose the rational subgroup:
• As the subgroup size increases, the control limits become
closer to central value, which makes the control chart more
sensitive.
• As the subgroup size increases, the inspection cost per
subgroup increases.
• When destructive testing is used and the item is expensive, a
small subgroup size of 2 or 3 is preferred.
• On the basis of statistical distribution, subgroup averages
become normal when subgroup size equals or exceeds 4
even the subgroups are taken from non-normal population.
• When the subgroup size exceeds from 10 or 12, the range (R
chart) should be replaced by standard deviation (S chart).

Industrial Engg. & Management, 1/e, 6

Pravin kumar, Pearson Education
Chart Techniques (Cont…)
• Collect the data
• Determine the trial central line and control limits
• Establish the revised central line and control limits
• Achieve the purpose

Industrial Engg. & Management, 1/e, 7

Pravin kumar, Pearson Education
Control Charts for Variables
• Control charts for variables monitor characteristics that can be measured and
have a continuous scale, such as height, weight, volume, or width.
• Objectives of Variable Control Charts
 There are following objectives of Variable Control Charts:
 To achieve quality improvement in products.
 To determine the process capability.
 To provide information for decision with regard to product specification.
 To provide information about current production process.
 To provide information about the decisions taken for currently produced
items.

Industrial Engg. & Management, 1/e, 8

Pravin kumar, Pearson Education
X-bar and R Chart
X 1  X 2  ...  X n R  X max.  X min.
X
n
g g

X j R1  R2  ...  Rg R j
X 1  X 2  ...  X g j 1 R 
j 1
X  g g
g g

UCLX  X  z X  X  3 X  X  A2 R
LCLX  X  z X  X  3 X  X  A2 R
3 ' 3
Here, 3 X   .R  A2 R
n d2 n

Industrial Engg. & Management, 1/e, 9

Pravin kumar, Pearson Education
Industrial Engg. & Management, 1/e, 10
Pravin kumar, Pearson Education
Industrial Engg. & Management, 1/e, 11
Pravin kumar, Pearson Education
 X-bar and R Chart
(Cont…) R 3d
UCLR R  3 R R  3d 3 R  3d 3 (1  3 ) R D4 R
d2 d2
R 3d 3
LCLR R  3 R R  3d 3 R  3d 3 (1  ) R D3 R
d2 d2

Establish the Revised control Limits

g g

X 
j 1
Xd R R
j 1
d

X0  ; R0 
g  gd g  gd
R0
UCLX  X 0  A 0  X 0  A
d2
R0
LCLX  X 0  A 0  X 0  A
d2
and UCLR D2 0
LCLR D1 0
Where the value of A, D1 , and D2 are the factors which will be take from Statistical Table Appendix.
Industrial Engg. & Management, 1/e, 12
Pravin kumar, Pearson Education
X-bar and R Chart
(Cont…)

Industrial Engg. & Management, 1/e, 13

Pravin kumar, Pearson Education
Example 1: A workshop produced the following data (in
millimeter) regarding a product manufactured on 5th August
2014 during two shifts of six hours each as shown in Table
below. Find the control limits for X-bar and R chart.
Subgr X1 X2 X3 X4 X- R Subg X1 X2 X3 X4 X- R
oup bar roup bar
No No
1 36 41 46 37 40 10 13 45 42 38 39 41 7
2 34 36 34 40 36 6 14 43 45 50 42 45 8
3 68 59 69 64 65 10 15 39 29 35 33 34 10
4 44 40 38 34 39 10 16 34 29 40 41 36 12
5 34 43 41 42 40 9 17 58 28 44 38 42 30
6 46 41 41 44 43 5 18 38 37 32 33 35 6
7 36 38 41 33 37 8 19 48 45 55 56 51 11
8 51 49 52 48 50 4 20 45 37 40 38 40 8
9 42 36 43 47 42 11 21 35 40 42 39 39 7
10 39 38 38 41 39 3 22 39 36 39 42 39 6
11 41 37 37 37 38 4 23 35 38 36 43 38 8
12 47 35 40 38 40 12 24 43 44 38 39 41 6

Industrial Engg. & Management, 1/e, 14

Pravin kumar, Pearson Education
Solution:
g

R
g

X 1  X 2  ...  X g
X j
990 R1  R2  ...  Rg j
211
j 1
X  
j 1
  41.25; R   8.79
g g 24 g g 24

UCLX  X  z X  X  3 X  X  A2 R  41.25  0.729 8.79  47.65

LCLX  X  z X  X  3 X  X  A2 R  41.25  0.729 8.79 34.84

Industrial Engg. & Management, 1/e, 15

Pravin kumar, Pearson Education
UCLR  D4 R  2.282 8.79  20.058
LCLR  D3 R  0 8.79  0

Industrial Engg. & Management, 1/e, 16

Pravin kumar, Pearson Education
Revised control limits
g

X  Xd
990  65  50  34  51
j 1
X0   39.5
g  gd 20
g

R R d
211  30
j 1
R0   10.73
g  gd 23
R0 10.73
UCLX  X 0  A 0  X 0  A 39.5  1.5   47.316
d2 2.059
R0 10.73
LCLX  X 0  A 0  X 0  A 39.5  1.5  31.683
d2 2.059
10.73
and UCLR  D2 0  4.698   24.482
2.059
10.73
LCLR  D1 0 0  0
2.059

Industrial Engg. & Management, 1/e, 17

Pravin kumar, Pearson Education
X-Bar & S-Chart
• For subgroup size less than 10, both R-chart and S chart shows the same variation
but for the subgroup size more than 10, S chart is preferred.

2
 n n

n X i    X i 
2

S j  i 1  i 1 
n n  1
g

X
j 1
j
X 1  X 2  ...  X n
X ; where X j  ; j
g n
g 2
n
 n  n
S n X    X i   X  X
2
j i i
S
j 1
; Where S j  i 1  i 1   i 1
; j
g n n  1 n  1
Industrial Engg. & Management, 1/e, 18
Pravin kumar, Pearson Education
 UCLX  X  A3 S ; LCLX  X  A3 S
UCLS B4 S ; LCLC B3 S

• Establish the Revised control Limits

g g

X 
j 1
Xd S  S
j 1
d
S0
X0  ; S0  ; 0 
g  gd g  gd c4
UCLX  X 0  A 0 ; LCLX  X 0  A 0
UCLS B6 0 ; LCLS B5 0

Industrial Engg. & Management, 1/e, 19

Pravin kumar, Pearson Education
X

Example 2: Find the control limits for X-bar and S chart from
the data given in Example 1.
Subgroup No X1 X2 X3 X4 s
1 36 41 46 37 40 4.546061
2 34 36 34 40 36 2.828427
3 68 59 69 64 65 4.546061
4 44 40 38 34 39 4.163332
5 34 43 41 42 40 4.082483
6 46 41 41 44 43 2.44949
7 36 38 41 33 37 3.366502
8 51 49 52 48 50 1.825742
9 42 36 43 47 42 4.546061
10 39 38 38 41 39 1.414214
11 41 37 37 37 38 2
12 47 35 40 38 40 5.09902
13 45 42 38 39 41 3.162278
14 43 45 50 42 45 3.559026
15 39 29 35 33 34 4.163332
16 34 29 40 41 36 5.597619
17 58 28 44 38 42 12.54326
18 38 37 32 33 35 2.94392
19 48 45 55 56 51 5.354126
20 45 37 40 38 40 3.559026
21 35 40 42 39 39 2.94392
22 39 36 39 42 39 2.44949
23 35 38 36 43 38 3.559026
24 43 44 38 39 41 2.94392

Industrial Engg. & Management, 1/e, 20

Pravin kumar, Pearson Education
Solution
g

X 1  X 2  ...  X g X j
990
j 1
X    41.25
g g 24

S 93.624
j 1
S0   3.901
g 24
UCLX  X  A3 S 41.25  1.628 3.901  47.6
LCLX  X  A3 S 41.25  1.628 3.901  34.899
UCLS  B4 S 2.266 3.901
LCLs  B3 S 0 3.901  0

Industrial Engg. & Management, 1/e, 21

Pravin kumar, Pearson Education
Industrial Engg. & Management, 1/e, 22
Pravin kumar, Pearson Education
Revised Control Chart
g g

X  Xd
990  65  50  34  51
S  S d
93.624  12.543
39.5;
j 1 j 1
X0   S0   3.252
g  gd 20 g  gd 23
S0 3.252
0   3.75
c4 0.94
UCLX  X 0  A 0 39.5  1.5 3.75  45.125; LCLX  X 0  A 0 39.5  1.5 3.75 33.875
UCLS  B6 0  2.088 3.75 7.83; LCLS  B5 0 0 3.75 0

Industrial Engg. & Management, 1/e, 23

Pravin kumar, Pearson Education
Control Chart for Trends

 X  W   W  W . X 
2

a
g  W   W 
2 2

g  W . X    W  X
b Slope of line 
g  W   W 
2 2

W = Subgroup numbers
represents the horizontal line
a = Intercepts of vertical axis

Industrial Engg. & Management, 1/e, 24

Pravin kumar, Pearson Education
Process Under Control
• About 2/3 (67%) of the points lie near the center line.
• A few points (5%) lie closer to control limits.
• The points are located back and forth across the center line.
• Points are balanced on both sides of the center line.
• No point lies beyond the control limits.

Industrial Engg. & Management, 1/e, 25

Pravin kumar, Pearson Education
Process Out of Control

Industrial Engg. & Management, 1/e, 26

Pravin kumar, Pearson Education
Reasons for process out of
control

• Change or Jump in level

• Trends or steady change in Level
• Recurring cycles
• Two Populations
• Mistakes

Industrial Engg. & Management, 1/e, 27

Pravin kumar, Pearson Education
 Process Capability and
Specification Limits
Central Limit Theorem
If the population from which samples are taken is not normal, the distribution of
sample average will tend towards normality provided that the sample size ‘n’ is at
least 4. This tendency increases as the sample size increases.

X - X -
z 
X 
n

Industrial Engg. & Management, 1/e, 28

Pravin kumar, Pearson Education
Process Capability and
Specification Limits (Cont…)
• Control Limits and Specifications

Industrial Engg. & Management, 1/e, 29

Pravin kumar, Pearson Education
Process Capability
• Take 20 subgroups of size 4 for a total 80 measurements. These
subgroups should be selected randomly to avoid any biasing.
• Calculate the sample standard deviation ‘S’ for each subgroup.
• Calculate the average sample standard deviation.

S=
 S S
=
g 20
• Estimate the population standard deviation, S
• Process capability is equal to 6 σ0 =
c4
0
• Capability Index,
U L
Cp  ;
6 0
where C p is capability index,U-L is tolerance,
and 6σ 0 is process capability.
Industrial Engg. & Management, 1/e, 30
Pravin kumar, Pearson Education
Example 23.3: Using the data provided in Example 1, find
the process capability index. The upper and lower specific
index is given as 50 mm and 30 mm.

Solution: Given U = 50 mm and L = 30 mm. From Example 1,

we find that the mean of standard deviation is 3.90 mm and
for subgroup size of 4, c4 = 0.9213.

S 3.90
σ0 =  4.23
c 4 0.9213

U  L 50  30
Cp   0.788
6 0 6 4.23
i.e. U  L  6 0

Since process spread greater than the difference between

specification limits is undesirable since, in this situation some
defective items may be produced.

Industrial Engg. & Management, 1/e, 31

Pravin kumar, Pearson Education
Limitations of chart for
Variables
• It cannot be used for the attributes type quality characteristics.
• If a product has large number of variables, it will require individual X-bar & R
chart for each variable which is not possible. In this case attributes chart is
preferred.

Industrial Engg. & Management, 1/e, 32

Pravin kumar, Pearson Education
THANKS