Unit 5

5.4 Statistical Quality

Control
 Quality control
 Quality control
 a process that evaluates output relative to a standard, and takes
corrective action when output doesn’t meet standards
 Purpose
 to assure that processes are performing in an
acceptable manner
 How
 monitoring process output using statistical
techniques
2
 Phases of Quality Assurance
Inspection and
Inspection corrective Quality built
before/after action during into the
production production process

Acceptance Process Continuous

sampling control improvement

The least The most

progressive progressive

3
 Quality control
 Quality control that relies primarily on inspection of
previously produced items is referred to as
acceptance sampling
 Quality control efforts that occur during production
are referred to as statistical process control
 The best companies emphasize designing quality
into the process (mistake prevention), thereby
greatly reducing the need for inspection or control
efforts
4
 Inspection
 Involves examining items to see if an item is good or defective
 The purpose of inspection is to provide information on the degree to
which items conform to a standard
 Inspection can occur before, during, or after production
 Inspection before and after production often involve acceptance
sampling procedures; monitoring during the production process is
referred to as process control

Inputs Transformation Outputs

Acceptance Process Acceptance

sampling control sampling
5
 Inspection
Basic issues
1. How Much/How Often
2. Where/When
3. Centralized vs. On-site
4. Whether to inspect Variables or Attributes

6
 How Much/How Often to
inspect
 Inspection may not be necessary for low value, high volume
products like common items such as pins, erasers or pencils
 Automated inspection may be necessary for high value, low
volume items
 The frequency of inspection depends largely on the rate at
which a process may go out of control or the number of lots
being inspected
 A stable process will require only infrequent checks, whereas
an unstable or one that has recently given trouble will require
more frequent checks
7
 Where/When to inspect in the
process
1. Raw materials and purchased parts
 Do not purchase poor quality products
2. Finished products
 Poor products returned by customers can also lead to additional
shipping costs
3. Before a costly operation
 not to waste Resources: labor, Material and Machine
4. Before an irreversible process
 Pottery, Ceramics, Tiles, PC chips, glass filaments
5. Before a covering process
 Before painting, plating and assembly-b/s they mask defects
8
 When to inspect in
- With increase in
inspection activities,
the cost of
the process undetected
defectives decreases
- With increase in
The amount of inspection is optimal when
inspection activities, the
the sum of the costs of inspection and
cost of inspection
passing defectives is minimized
increases

Total Cost
Cost

Cost of
inspection
Cost of
passing
defectives
Optimal

Amount of Inspection 9
 Examples of Inspection Points
– service organizations
Type of Inspection Characteristics
business points
Fast Food Cashier Accuracy
Counter area Appearance, productivity
Eating area Cleanliness
Building Appearance
Kitchen Health regulations
Hotel/motel Parking lot Safe, well lighted
Accounting Accuracy, timeliness
Building Appearance, safety
Main desk Waiting times
Supermarket Cashiers Accuracy, courtesy
Deliveries Quality, quantity
Centralized VS On-site
Inspection
 On-sit inspection
 Ships, Nuclear Plants, Petroleum Refinery,
Chemical Plant equipments etc both external and
internal inspection
 Centralized inspection
 specialized tests in the lab such as blood tests,
analyzing food samples

11
Statistical Process Control

 All work occurs in a system of

interconnected processes
 Variation exists in all processes
 Understanding and reducing variation
are the keys to success

12
 Statistical Process Control
(SPC)

 Variability is inherent in every process

 Natural or Common causes of variation

 Special or Assignable causes of variation

 SPC provides a statistical signal when assignable

causes are present

13
 Natural Variations
 Also called common causes/or random variation/
 Affect virtually all production processes
 Expected amount of variation
 Are likely to occur in a random manner and about which the firm can do
little
 The presence of these causes in the system is due to a multitude of
reasons which are difficult to identify and uneconomical to eliminate
 They occur randomly and can be described by the normal probability
distribution
 For any distribution there is a measure of central tendency and
dispersion
 If the distribution of outputs falls within acceptable limits, the process is
said to be “in control”
14
 Assignable Variations
 Also called special causes of variation /or non-
random variation/
 Generally this is some change in the process
 Variations which occur less frequently and can
normally be traced to a specific reason
 The objective is to discover when assignable causes
are present
 Eliminate the bad causes
 Incorporate the good causes
15
 Statistical Process Control
(SPC)
 The essence of SPC is to assure that the output of a
process is random so that future output will be random
 Basic assumptions of SPC:
 Every process has random /natural/ variation in it
 Production processes are not usually found in a state of control
 “State of Control”
 Unnecessary variation is eliminated
 Remaining variation is because of random/natural / causes
 Main Task of process control: is to distinguish random
from non-random variability, because non random
variability indicates that the process is out of control
16
 Process Control Chart Statistical process control
(SPC): involves testing a
random sample of output
from a process to
determine whether the
 Control Chart process is producing items
within a pre-selected range
 Purpose: to monitor process output to see if it is
random
 A time ordered plot of representative sample
statistics obtained from an on going process
(e.g. sample means)
 Upper and lower control limits define the range
of acceptable variation
17
 Control Chart
Abnormal variation Out of
due to assignable sources control
UCL

Mean
Normal variation
due to chance
LCL
Abnormal variation
due to assignable sources

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sample number

18
 Statistical Process Control
 Variations and Control
 Random variation: Natural variations in the
output of a process, created by countless
minor factors
 Assignable variation: A variation whose
source can be identified

19
Statistical Process Control

If only natural causes of variation are present, the

output of a process forms a distribution that is
stable over time and is predictable

Prediction

Figure S6.1

20
Statistical Process Control
If assignable causes are present, the process output is
not stable over time and is not predicable

?
?? ??
? ?
? ?
? ?
? ?
?? ??
? ?

Prediction

Figure S6.1

21
Control Charts
the purpose of control charts is to help distinguish
between natural variations and variations due to
assignable causes
A necessary (but not sufficient) condition for a process
to be deemed “in control” or stable, is for all the data
points to fall b/n the upper and lower control limits

22
Sampling Distribution

Sampling
distribution

Process
distribution

Mean

23
 Normal Distribution
Percentage of values with in given ranges in a normal distribution

Standard deviation

   

Mean
95.44%

99.74%

24
Control Limits
Sampling
distribution

Process
distribution

Mean

Lower Upper
control control
limit limit 25
 SPC Errors (basic management
mistake?)
 Type I error
 Concluding a process is not in control when it actually is.
 Treating as a special cause any fault, complaint, mistake,
breakdown, accident or shortage when it actually is due to
common causes
 Type II error

Concluding a process is in control when it is not .
 Attributing to common causes any fault, complaint,
mistake, breakdown, accident or shortage when it actually
is due to a special cause
26
 Type I Error

/2 /2

Mean

Probability LCL UCL

of Type I error

27
 Control Charts for Variables
 Mean control charts
 Used to monitor the central tendency of a
process.
 X bar charts
 Range control charts
 Used to monitor the process dispersion
 R charts
28
Example of x-Bar and R Charts:
Required Data
Sample Obs 1 Obs 2 Obs 3 Obs 4 Obs 5
1 10.682 10.689 10.776 10.798 10.714
2 10.787 10.86 10.601 10.746 10.779
3 10.78 10.667 10.838 10.785 10.723
4 10.591 10.727 10.812 10.775 10.73
5 10.693 10.708 10.79 10.758 10.671
6 10.749 10.714 10.738 10.719 10.606
7 10.791 10.713 10.689 10.877 10.603
8 10.744 10.779 10.11 10.737 10.75
9 10.769 10.773 10.641 10.644 10.725
10 10.718 10.671 10.708 10.85 10.712
11 10.787 10.821 10.764 10.658 10.708
12 10.622 10.802 10.818 10.872 10.727
13 10.657 10.822 10.893 10.544 10.75
14 10.806 10.749 10.859 10.801 10.701
15 10.66 10.681 10.644 10.747 10.728
29
Example of x-bar and R charts: Step 1. Calculate
sample means, sample ranges, mean of means, and
mean of ranges.
Sample Obs 1 Obs 2 Obs 3 Obs 4 Obs 5 Avg Range
1 10.682 10.689 10.776 10.798 10.714 10.732 0.116
2 10.787 10.86 10.601 10.746 10.779 10.755 0.259
3 10.780 10.667 10.838 10.785 10.723 10.759 0.171
4 10.591 10.727 10.812 10.775 10.73 10.727 0.221
5 10.693 10.708 10.79 10.758 10.671 10.724 0.119
6 10.749 10.714 10.738 10.719 10.606 10.705 0.143
7 10.791 10.713 10.689 10.877 10.603 10.735 0.274
8 10.744 10.779 10.11 10.737 10.75 10.624 0.669
9 10.769 10.773 10.641 10.644 10.725 10.710 0.132
10 10.718 10.671 10.708 10.85 10.712 10.732 0.179
11 10.787 10.821 10.764 10.658 10.708 10.748 0.163
12 10.622 10.802 10.818 10.872 10.727 10.768 0.250
13 10.657 10.822 10.893 10.544 10.75 10.733 0.349
14 10.806 10.749 10.859 10.801 10.701 10.783 0.158
15 10.660 10.681 10.644 10.747 10.728 10.692 0.103

Averages 10.728 0.220400

30
Example of x-bar and R charts: Step 2.
Determine Control Limit Formulas and
Necessary Tabled Values

x Chart Control Limits

n A2 D3 D4
UCL = x + A 2 R 2 1.88 0 3.27
3 1.02 0 2.57
LCL = x - A 2 R 4 0.73 0 2.28
5 0.58 0 2.11
6 0.48 0 2.00
R Chart Control Limits 7 0.42 0.08 1.92
8 0.37 0.14 1.86
UCL = D 4 R 9 0.34 0.18 1.82
10 0.31 0.22 1.78
LCL = D 3 R 11 0.29 0.26 1.74

31
Example of x-bar and R charts: Steps 3&4.
Calculate x-bar Chart and Plot Values

UCL = x + A 2 R  10.728  .58(0.2204) = 10.856

LCL = x - A 2 R  10.728 - .58(0.2204) = 10.601
10.900

10.850 UCL
10.800

10.750
M eans

10.700

10.650

10.600
LCL
10.550
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Sam ple
32
Example of x-bar and R charts: Steps 5&6.
Calculate R-chart and Plot Values

UCL = D 4 R  ( 2.11)( 0.2204 )  0.46504

LCL = D 3 R  ( 0)( 0.2204 )  0
0 .8 0 0

0 .7 0 0

0 .6 0 0

0 .5 0 0
UCL
R 0 .4 0 0

0 .3 0 0

0 .2 0 0

0 .1 0 0

0 .0 0 0
LCL
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
S a mp le
33
 Mean and Range Charts
(process mean is
shifting upward)
Sampling
Distribution

UCL

x-Chart Detects shift

LCL

UCL

Does not
R-chart
detect shift
LCL
 Mean and Range Charts
Sampling
Distribution (process variability is increasing)

UCL

Does not
x-Chart
LCL
reveal increase

UCL

R-chart Reveals increase

LCL

35
 Control Chart for Attributes
 p-Chart - Control chart used to monitor
the proportion of defectives in a process
 c-Chart - Control chart used to monitor
the number of defects per unit of
sample

36
Use of p-Charts
 When observations can be placed into
two categories i.e., mutually exclusive
 Good or bad
 Pass or fail
 Operate or don’t operate
 When the data consists of multiple
samples of several observations each

37
 Use of c-Charts
 Use only when the number of occurrences
per unit of measure can be counted; non-
occurrences cannot be counted.
 Scratches, chips, dents, or errors per item
 Cracks or faults per unit of distance
 Breaks or Tears per unit of area
 Bacteria or pollutants per unit of volume
 Calls, complaints, failures per unit of time

38
Example of Constructing a p-Chart:
Required Data
1 100 4
2 100 2
3 100 5
4 100 3
5 100 6
6 100 4
7 100 3
8 100 7
9 100 1
10 100 2
11 100 3
12 100 2
13 100 2
14 100 8
15 100 3
39
Statistical Process Control Formulas:
Attribute Measurements (p-Chart)

Given: Total Number of Defectives

p=
Total Number of Observations
p (1 - p)
sp =
n

Compute control limits:

UCL = p + z sp
LCL = p - z sp
40
Example of Constructing a p-chart: Step 1
Sample n Defectives p
1 100 4 0.04
1. Calculate the sample 2 100 2 0.02
3 100 5 0.05
proportions, p (these 4 100 3 0.03
are what can be plotted 5 100 6 0.06
on the p-chart) for each 6 100 4 0.04
7 100 3 0.03
sample.
8 100 7 0.07
9 100 1 0.01
10 100 2 0.02
11 100 3 0.03
12 100 2 0.02
13 100 2 0.02
14 100 8 0.08
15 100 3 0.03
41
Example of Constructing a p-chart:
Steps 2&3

2. Calculate the average of the sample proportions.

55
p= = 0.036
1500
3. Calculate the standard deviation of the sample
proportion

p (1 - p) .036(1 - .036)
sp = = = .0188
n 100
42
Example of Constructing a p-chart: Step 4

4. Calculate the control limits.

UCL = p + z sp
LCL = p - z sp

.036  3(.0188)

UCL = 0.0924
LCL = -0.0204 (or 0)

43
Example of Constructing a p-Chart: Step 5
Sample n Defectives p
1 100 4 0.04
5. Plot the individual sample proportions, the 2 100 2 0.02
average of the proportions, and the control limits 3 100 5 0.05
4 100 3 0.03
5 100 6 0.06
0.16
6 100 4 0.04
0.14 7 100 3 0.03
8 100 7 0.07
0.12 9 100 1 0.01
10 100 2 0.02
0.1 11 100 UCL
3 0.03
12 100 2 0.02
p 0.08
13 100 2 0.02
14 100 8 0.08
0.06
15 100 3 0.03
0.04

0.02

0 LCL
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Observation 44
 Control Limits for c-Charts
Population will be a Poisson distribution, but
applying the Central Limit Theorem allows us
to assume a normal distribution for the sample
statistics
c  c UCL c  c  z c LCLc  c  z c

UCLc = c + 3 c LCLc = c - 3 c

c =mean number defective in the sample

number of incidents observed

c
number of units sampled 45
 Example: c-Chart for Cab
Company
c = 54 complaints/9 days = 6 complaints/day

UCLc = c + 3 c 14 –
UCLc = 13.35

Number defective
=6+3 6 12 –
= 13.35 10 –
8 –
6 – c= 6
LCLc = c - 3 c 4 –
=6-3 6 2 – LCLc = 0
0 – | | | | | | | | |
=0 1 2 3 4 5 6 7 8 9
Day

46
Patterns in Control Charts
Upper control limit

Target

Lower control limit

Normal behavior. Process is “in

control.”

47
Patterns in Control Charts
Upper control limit

Target

Lower control limit

One plot out above (or below).
Investigate for cause. Process is
“out of control.”

48
Patterns in Control Charts
Upper control limit

Target

Lower control limit

Trends in either direction, 5 plots.
Investigate for cause of
progressive change.

49
Patterns in Control Charts
Upper control limit

Target

Lower control limit

Two plots very near lower (or
upper) control. Investigate for
cause.

50
Patterns in Control Charts
Upper control limit

Target

Lower control limit

Run of 5 above (or below) central
line. Investigate for cause.

51
Patterns in Control Charts
Upper control limit

Target

Lower control limit

Erratic behavior. Investigate.

52
 Run Tests
 Run test – a test for randomness
 Any sort of pattern in the data would
suggest a non-random process
 All points are within the control limits -
the process may not be random

53
 Nonrandom Patterns in Control
charts

 Trend
 Cycles
 Bias
 Mean shift
 Too much dispersion

54
Counting Runs
Counting Above/Below Median Runs (7 runs)

B A A B A B B B A A B

Counting Up/Down Runs (8 runs)

55
U U D U D U D U U D
 Process Capability
 Tolerances or specifications
 Range of acceptable values established by
engineering design or customer requirements
 Process variability
 Natural variability in a process
 Measured in terms of process standard deviation
 Process capability
 Process variability relative to the variation allowed by
specification 56
 Process Variability
Lower Upper
Specification Specification

A. Process variability
matches specifications
Lower Upper
Specification Specification

B. Process variability
well within specifications
Upper
Specification
Lower
Specification
C. Process variability 57
exceeds specifications
 Process Capability Ratio

specification width ≥ 1
Process capability ratio, Cp = process width

Upper specification – lower specification

Cp =
6

58
 - For a process to be deemed to be capable, it must
have a capability ratio of at least 1.33
3 Sigma and 6 - A ratio of 1.33 implies only about 30 parts per
million (ppm) can be expected to not be within the
Sigma Quality specifications
- the greater the capability ratio, the greater the
probability that the output of a machine or process
will fall within design specifications
Lower
specification
Upper
1350 ppm specification
1350 ppm

1.7 ppm 1.7 ppm

Process
mean
+/- 3 Sigma

+/- 6 Sigma
59
Process Capability Ratio
Upper Specification - Lower Specification
Cp =
6
 A capable process must have a Cp of at
least 1.0
 Does not look at how well the process is
centered in the specification range
 Often a target value of Cp = 1.33 is used
to allow for off-center processes
 Six Sigma quality requires a Cp = 2.0 60
Process Capability Ratio
Insurance claims process
Process mean x = 210.0 minutes
Process standard deviation  = .516 minutes
Design specification = 210 ± 3 minutes

Upper Specification - Lower Specification

Cp = 6

61
Process Capability Ratio
Insurance claims process
Process mean x = 210.0 minutes
Process standard deviation  = .516 minutes
Design specification = 210 ± 3 minutes

Upper Specification - Lower Specification

Cp = 6

213 - 207
= 6(.516)
= 1.938 Process is
capable
62
If the process capability of a normally
distributed process is .084, the process is in
control, and is centered at .550. What are the
upper and lower control limits for this process?

Process Capability = 6 

6 = .084 = .014

UCL = .550 + 3(.014)= .592

LCL = .550 - 3(.014) = .508

63
Process Capability Chart
Process output
distribution

Output Output
out of spec out of spec

5.010

4.90 4.95 5.00 5.05 5.10 5.15

cm
X
Tolerance band
LSL USL
Process capability (6s )
64
 Process Capability
This process is
CAPABLE of
producing all good
output.

ä Control the process.

Lower Upper
Specification Specification
Limit Limit This process is
× NOT CAPABLE.

ä INSPECT - Sort out

the defectives
65
Process Capability Index, Cpk

Capability Index shows  X  LTL UTL - X 

C pk 
= min  or 
3  3 
how well parts being  
produced fit into design
limit specifications.

As a production
process produces
items small shifts in
equipment or systems
can cause differences
in production
performance from
differing samples.
Shifts in Process Mean
66
 Process Capability Index
Upper Lower
Specification - x x - Specification
Cpk = minimum of Limit , Limit
 

 A capable process must have a Cpk of at least

1.0
 A capable process is not necessarily in the
center of the specification, but it falls within
the specification limit at both extremes
67
Process Capability Index

New Cutting Machine

New process mean x = .250 inches
Process standard deviation  = .0005 inches
Upper Specification Limit = .255 inches
Lower Specification Limit = .249 inches

68
Process Capability Index
New Cutting Machine

New process mean x = .250 inches

Process standard deviation  = .0005 inches
Upper Specification Limit = .255 inches
Lower Specification Limit = .249 inches

(.255) - .250
Cpk = minimum of ,
(3).0005

69
Process Capability Index
New Cutting Machine
New process mean x = .250 inches
Process standard deviation  = .0005 inches
Upper Specification Limit = .255 inches
Lower Specification Limit = .249 inches

(.255) - .250 .250 - (.249)

Cpk = minimum of ,
(3).0005 (3).0005
Cpk=0.005/0.0015= 3.333 or

.001
New machine is
Cpk =
.0015
= 0.67 NOT capable
70
 Process Capability Index- Example

Given: Smaller of:

{
 process mean = –
Upper Tol Limit - X
1.0015 3
=
 = .001 Cpk OR –
X - Lower Tol Limit
 LTL = .994 3
 UTL = 1.006

1.0015 -.994 1.006 - 1.0015

or
Cpk= min
3(.001) 3 (.001)

Cpk= min [2.5 or 1.5] = 1.5

71
 Process Capability: Cpk Varieties

Cpk = 1.0 Cpk = 1.33 Cpk = 3.0

LSL USL LSL USL LSL USL

(d) (f)

Cpk = 1.0 Cpk = 0.60 Cpk = 0.80

LSL USL LTL USL LTL USL

72
 Improving Process Capability
 Simplify
 Standardize
 Mistake-proof
 Upgrade equipment
 Automate

73
Acceptance
Acceptance Sampling
Sampling
 Acceptance Sampling
 A statistical quality control technique used
in deciding to accept or reject a shipment
of input or output.
 Acceptance sampling inspection can range
from 100% of the Lot to a relatively few
items from the Lot (N=2) from which the
receiving firm draws inferences about the
whole shipment.
74
 Acceptance Sampling
 Purposes
 Determine quality level
 Ensure quality is within predetermined level
 Advantages
 Economy
 Less handling damage
 Fewer inspectors
 Upgrading of the inspection job
 Applicability to destructive testing
 Entire lot rejection (motivation for improvement)
75
Acceptance Sampling

 Disadvantages
 Risks of accepting “bad” lots and rejecting
“good” lots
 Added planning and documentation
 Sample provides less information than 100-
percent inspection

76
Statistical
Statistical Sampling
Sampling
Techniques
Techniques
 n and c
 The bottom line in acceptance sampling is
that acceptance sampling plans are
designed to give us two things: n and c,
where
n = the sample size of a particular sampling plan
c = the maximum number of defective pieces for
a sample to be rejected

78
END

79