You are on page 1of 68

Statistical Process

Control
Shashanka RH
Faculty, HMA
9986192247
shashanka.rh@gmail.com

No. 34329
Agenda
• Introduction to SQC/SPC and its historical perspective.
• What is Data?
• Types of Data
• Descriptive Statistics
• 7 Quality tools
• Process Stability
• Process Capability
• How to use MS Excel for SPC studies
• Basics of sampling
HAL Management Academy 2
History
• SQC was pioneered by Walter A. Shewhart at Bell Laboratories in
the early 1920s.
• Shewhart developed the control chart in 1924 and the concept of a
state of statistical control.
• Shewhart consulted with Colonel Leslie E. Simon in the application
of control charts to munitions manufacture at the Army's
Picatinney Arsenal in 1934.

3
History
• W. Edwards Deming invited Shewhart to speak at the Graduate
School of the U.S. Department of Agriculture, and served as the
editor of Shewhart's book Statistical Method from the Viewpoint of
Quality Control (1939) which was the result of that lecture.
• Deming was an important architect of the quality control short
courses that trained American industry in the new techniques
during WWII.
• Deming traveled to Japan during the Allied Occupation and met
with the Union of Japanese Scientists and Engineers(JUSE)in an
effort to introduce SQC methods to Japanese industry

4
What is SQC ?
• Statistical quality control (SQC) is the term used to describe the
set of statistical tools used by quality professionals.

5
DATA GATHERING
• What is Data ?
Data is a numerical expression of an activity.

Conclusions based on facts and data are necessary for any improvement.

-K. Ishikawa

If you are not able to express a phenomenon in numbers, you do not know about it
adequately.

-Lord Kelvin

6
TYPES OF DATA

CONTINUOUS DISCRETE
Measurable Subjective Assessment
e.g. :Length, Temperature e.g. :Score in a beauty
contest
Countable
e.g. :Number of defects

7
WHAT IS THE DIFFERENCE
THE NUMBER OF SHAFTS
BETWEEN
A SHAFT DIAMETER
REJECTED FOR OVERSIZE
DIAMETER
The diameter of a shaft can The number of shaft rejected has
take any value ever after the necessarily to be a whole number.
e.g.. 0, 2, 7, 10 numbers rejected
decimal point e.g.. 19.055,
etc..
19.0516 etc..
Data related to this type of
Data related to this type of
parameters are called Discrete
parameters are called
data.
Continuous data.

8
WHICH OF THE BELOW ARE
CONTINUOUS / DISCRETE DATA?
• Width of sheet
• No. of liners thinned
• Tubes rejected by Go- Nogo Gauge
• Diameter of Piston
• Height of a Man
• Sheet thickness
• Out of 100 sheets the numbers that meet the thickness 4  0.9
• Time taken to process a purchase order
• No. of bugs in a program

9
Types of Data
Data

Discrete Continuous

Binary Unordered Ordered Count


categories categories

Classified No rankings Rankings Counted Measured


Description into one of or ratings discretely on a
two categories continuum
or scale

Example
% of Customer Number of Time to
Types of
applications satisfaction errors in an process an
products
with one or rating of application application
more errors call center
service

More Info.
10
OBJECTIVES OF DATA COLLECTION
• To know and quantify the status
• To monitor the process
• To decide acceptance or rejection
• To analyse and decide the course of action

HOW TO COLLECT DATA ?


• Define the purpose
• Decide the type of analysis
• Define the period of data collection
• Is the required data already available ?

11
SQC Categories
Statistical Quality Control

Descriptive Statistics

Statistical Process Control (SPC)

Sampling

12
Descriptive statistics
Descriptive Statistics

Central Tendency

Mean

Median

Mode

Dispersion

Range

Variance

Standard Deviation

13
Central Tendency
• The Mean is the average or arithmetic mean of the data.
• The Median is the middle value of the data set, when the dataset is
arranged in the ascending or descending order.
• The Mode is the most frequently observed value(s).

Dispersion
• The Range is the difference between highest and lowest
observations in the dataset.
• The Variance is a measure of how spread out a data set is. It is
calculated as the average squared deviation of each number from
the mean of a data set.
• The Standard Deviation is the Square Root of the Variance.
14
n
Mean x i
i 1
X
n Note: These formulae are for Population
parameters.

2 For Standard Deviation and Variance of a

Variance 
n
( x   ) 2
sample, use “n-1” instead of “n” in the
i
 , formulae.

i 1 n Remember that the formula for “mean”


remains unchanged. This is because “Mean”
is an unbiased estimator.

2
Standard  ( xn
  )
i

Deviation i 1 n
15
Spot Quiz 2
1. Find the mean, median and mode of the following dataset:
24, 32, 15, 26, 20, 16, 28, 30, 17, 27
a) 23.5, 25, none
b) 24, 18, none
c) 23.5, 25, all the above
d) 24, 18, all the above

2. Find the mode of the following dataset:


24, 32, 15, 15, 20, 16, 28, 30, 17, 17
e) 15
f) 17
g) 15 and 17
h) None, since there cannot be more than one mode

Shashanka RH, HMA 16


For a part with required diameter of 36 mm,
Machine-A produces parts with mean_A = 35mm and SD_A = 1.5mm.
Machine-B produces parts with mean_B = 32mm and SD_A = 0.8mm.
Which machine is better and which is more consistent?

a) A is better, B is consistent
b) B is better, B is consistent
c) B is better, A is consistent
d) A is better, A is consistent
e) Such a comparison cannot be done

Shashanka RH, HMA 17


SQC Categories
Statistical Quality Control

Descriptive Statistics

Statistical Process Control (SPC)

Sampling

18
Statistical process control
• Extend the use of descriptive statistics to monitor the quality of the
product and process.
• Statistical process control help to determine the amount of variation to
make sure the process is in a state of control

Ishikawa’s Basic Tools of Quality


• Kaoru Ishikawa developed seven basic visual tools of quality so that the
average person could analyze and interpret data.
• These tools have been used worldwide by companies, managers of all
levels and employees.

19
Seven Quality Tools
Histograms
• Histogram Defined
• A histogram is a bar graph that shows frequency data.
• Histograms provide the easiest way to evaluate the distribution of data.
Histograms
• Creating a Histogram
• Collect data and sort it into categories.
• Then label the data as the independent set or the dependent set.
• The characteristic you grouped the data by would be the independent variable.
• The frequency of that set would be the dependent variable.
• Each mark on either axis should be in equal increments.
• For each category, find the related frequency and make the horizontal marks
to show that frequency.
Pareto Charts
• Pareto Chart Defined
• Pareto charts are used to identify and prioritize problems to be solved.
• They are actually histograms aided by the 80/20 rule adapted by Joseph
Juran.
• Remember the 80/20 rule states that approximately 80% of the problems are created by
approximately 20% of the causes.
Pareto Charts
• Constructing a Pareto Chart
• First, information must be selected based on types or classifications of defects
that occur as a result of a process.
• The data must be collected and classified into categories.
• Then a histogram or frequency chart is constructed showing the number of
occurrences.
Pareto Charts
• An Example of How a Pareto Chart Can Be Used
• Pareto Charts are used when products are suffering from different defects but the defects are
occurring at a different frequency, or only a few account for most of the defects present, or
different defects incur different costs.
• What we see from that is a product line may experience a range of defects. The
manufacturer could concentrate on reducing the defects which make up a bigger percentage
of all the defects or focus on eliminating the defect that causes monetary loss.
Pareto Charts
Cause and Effect Diagrams
• Cause and Effect Diagram Defined
• The cause and effect diagram is also called the Ishikawa diagram or the fishbone diagram.
• It is a tool for discovering all the possible causes for a particular effect.
• The major purpose of this diagram is to act as a first step in problem solving by creating a list
of possible causes.
Cause and Effect Diagrams
• Constructing a Cause and Effect Diagram
• First, clearly identify and define the problem or effect for which the causes must be
identified. Place the problem or effect at the right or the head of the diagram.
• Identify all the broad areas of the problem.
• Write in all the detailed possible causes in each of the broad areas.
• Each cause identified should be looked upon for further more specific causes.
• View the diagram and evaluate the main causes.
• Set goals and take action on the main causes.
Cause and Effect Diagrams
• An Example of When a Cause and Effect Diagram Can Be Used
• This diagram can be used to detect the problem of incorrect deliveries.
• Diagram on next slide
• Diagram obtained from: <http://www.hci.com.au/hcisite/toolkit/causeand.htm>
• When a production team is about to launch a new product, the factors that
will affect the final product must be recognized. The fishbone diagram can
depict problems before they have a chance to begin.
Cause and Effect Diagrams
Scatter Diagrams
• Scatter Diagrams Defined
• Scatter Diagrams are used to study and identify the possible relationship
between the changes observed in two different sets of variables.
Scatter Diagrams
• Constructing a Scatter Diagram
• First, collect two pieces of data and create a summary table of the data.
• Draw a diagram labeling the horizontal and vertical axes.
• It is common that the “cause” variable be labeled on the X axis and the “effect” variable be labeled
on the Y axis.
• Plot the data pairs on the diagram.
• Interpret the scatter diagram for direction and strength.
Scatter Diagrams
• An Example of When a Scatter Diagram Can Be Used
• A scatter diagram can be used to identify the relationship between the
production speed of an operation and the number of defective parts made.
Flow Charts
• Flow Charts Defined
• A flow chart is a pictorial representation showing all of the steps of a process.
Flow Charts
• Creating a Flow Chart
• First, familiarize the participants with the flow chart symbols.
• Draw the process flow chart and fill it out in detail about each element.
• Analyze the flow chart. Determine which steps add value and which don’t in
the process of simplifying the work.
Run Charts
• Run Charts Defined
• Run charts are used to analyze processes according to time or order.
Run Charts
• Creating a Run Chart
• Gathering Data
• Some type of process or operation must be available to take measurements for analysis.
• Organizing Data
• Data must be divided into two sets of values X and Y. X values represent time and values of Y represent
the measurements taken from the manufacturing process or operation.
• Charting Data
• Plot the Y values versus the X values.
• Interpreting Data
• Interpret the data and draw any conclusions that will be beneficial to the process or operation.
Run Charts
• An Example of Using a Run Chart
• An organization’s desire is to have their product arrive to their customers on
time, but they have noticed that it doesn’t take the same amount of time
each day of the week. They decided to monitor the amount of time it takes to
deliver their product over the next few weeks.
Process Stability

• So, now we know whether the process is stable or not i.e., is it under
the influence of common or special causes

•This helps partly decide our improvement strategy

•Common Causes – need to make systemic changes

•Special Causes – investigate root causes to fix and eliminate special


causes

39
Control Charts
• Control Charts Defined
• Control charts are used to determine whether a process will produce a
product or service with consistent measurable properties.
Control Charts
• Steps Used in Developing Process Control Charts
• Identify critical operations in the process where inspection might be needed.
• Identify critical product characteristics.
• Determine whether the critical product characteristic is a variable or an
attribute.
• Select the appropriate process control chart.
• Establish the control limits and use the chart to monitor and improve.
• Update the limits.
Control Charts
• An Example of When to Use a Control Chart
• Counting the number of defective products or services
• Do you count the number of defects in a given product or service?
• Is the number of units checked or tested constant?
Variation in Quality
• No two items are exactly alike.
• Some sort of variations in the two items is bound to be there. In
fact it is an integral part of any manufacturing process.
• This difference in characteristics known as variation.
• This variation may be due to substandard quality of raw material,
carelessness on the part of operator, fault in machinery system
etc..

Copyright@ HMA 2021. All rights reserved.


43
This presentation is part of the online training modules offered by HAL Management Academy through LMS.
Variation in a process

Chance Causes Assignable Causes

Copyright@ HMA 2021. All rights reserved.


44
This presentation is part of the online training modules offered by HAL Management Academy through LMS.
Variation due to chance causes/common
causes
• Variation occurred due to chance.
•  This variation is NOT due to defect in machine, Raw
• material or any other factors.
•  Behave in “random manner”.
•  Negligible but Inevitable
•  The process is said to be under the state of statistical
• control.

45
Variation due to assignable
causes

• Non – random causes like:


• Difference in quality of raw material
• Difference in machines
• Difference in operators
• Difference of time

46
47
Specification and control limits
• No item in the world can be a true copy of another item.
• It is not expressed in absolute values but in terms of a range.
• For Eg:
• The diameter of a pen is expected by its
• manufacturer not as 7mm but as 7mm ± 0.05.
• Thus, the diameter of a pen produced by the
• manufacturer can vary from 6.95 mm to 7.05 mm.

48
Setting Control Limits

49
Interpreting Control Charts : Special
Pattern Plot Conclude Action Example

Shifts
Cause Patterns
8 or more points in a row on
the same side of the median
indicate a shift in a key element
Find out what was
different about the
process around the time
Customer
complaints
increase due to
change in policy
of the process. that the shift occurred.

Trends 7 or more points in a row Find out what was Market growth or
continuously increasing or different about the decline
continuously decreasing process around the time
indicate a trend. that the trend started.

Same A sequence of 7 or more Find out if measurement Cycle time


Value points having the same device is stuck. measured to
value. nearest day
Poor resolution.

Cycle 14 or more consecutive points Find out what was different Fluctuations in
up and down about the process around staffing due to
the time the cycle occurred holidays

50
Quick Breakout
Example 1 : Is there a Special Cause Pattern?
Activity

97

Median = 92
% Responding
Completely/Very
92
Satisfied

87

82June July August

51
Quick Breakout
Example 2 : Is there a Special Cause Pattern?
Activity

80

# of
Contracts 70 Median = 63.15
(000’s)

60

50

40
June July August

52
Types of Control Charts

Attribute Control Charts


Variable Control Charts
Process Capability
• Process Capability Studies are designed to see what the process is
“capable” of doing under controlled conditions.
• An “in-control” process can produce bad or out-of-spec product.
Manufacturing processes must meet or be able to achieve product
specifications.
• Process capability is the repeatability and consistency of a
manufacturing process relative to the customer requirements in
terms of specification limits of a product parameter.
• This measure is used to objectively measure the degree to which your
process is or is not meeting the requirements.

HAL Management Academy 54


IMPORTANT SPC RATIOS USED
LSL x USL

--3s
3s ++3s
3

Maximum allowable range of characteri


stic
Cp 
Normal variation of process
Tolerance USL - L SL
 
6 s 6 s
This compares the requirement of the process output vis-a-vis the inherent
variability of the process. Higher value than 1 implies that the process has got
the capability to give the product within the set limits.
55
PROCESS POTENTIAL INDEX (Cp)
The numerator is
controlled by Design
Engineering
Maximum Allowable Range of Characteristic
Cp =
Normal Variation of Process

The denominator is controlled by


Process Engineering

56
USL -X
Cpu 
3 s
This gives us the positioning of the mean vis-a-vis the USL and
the relationship between the two.
X  LSL
Cpl 
3 s
This gives us the positioning of the mean vis-a-vis the LSL and
the relationship between the two.
Cpk - Process Performance Index. This is important

Cpk = Minimum of (Cpu and Cpl) ; for bilateral tolerances


+Y
= C pu ;for unilateral tolerance on upper side i.e.. X -O

+O
= Cpl ;for unilateral tolerance on lower side i.e.. X -Y

57
These ratios help you in:

 Predicting whether rejections will take place on the higher


side or on the lower side.

 Taking centering decisions.

 Deciding whether to consider broadening of tolerances

 Taking Decisions on whether to go in for new m/cs.

 Deciding on the level of inspection required.

58
SOME TYPES
LSL USL LSL USL

Failure likely on higher side Failure likely on lower side

TOLERANCE OR NEW
CENTERING RELATED PROBLEMS MACHINE DECISION
LSL USL LSL USL

LSL USL
LSL USL LSL USL

59
SQC Categories
Statistical Quality Control

Descriptive Statistics

Statistical Process Control (SPC)

Sampling

60
POPULATION AND SAMPLE
• The entire set of items is called the Population.
• The small number of items taken from the population to
make a judgment of the population is called a Sample.
• The numbers of samples taken to make this judgment is
called Sample size.

POPULATION SAMPLE OF SIZE


THREE

61
REPRESENTATIVE SAMPLES

• For conclusions to be valid, samples must be representative.


– Data should fairly represent the population or process
– No systematic differences should exist between the data you collect
and the data you don’t collect
62
Sampling
Where Are You Standing?
Strategy
Process Data

Population Data

Population Approach
• Make probability statements about the population from the sample
» “I have 95% confidence that the mean of the population is between 1.5 and 2.5 seconds.”
• Use sample size formula

Process Approach
• Assess the stability of the population over time
» Are shifts, trends, or cycles occurring?
» Do I take a special or common cause variation approach to process improvement?
• Use rational subgrouping

63
Population Study
Sampling Strategy :
Random Sampling

Population Sample Description


N n

Each unit (“X”) has


XXX an equal probability
XXXXX X X X of being selected in
XXXXX
XXX
a sample

64
Sampling Strategy : Stratified
Population Study
Random Sampling
Population Sample Description

Segments Units Randomly sample


within a stratified category
Large LLLLL
L or group
MMM
Medium MMM
MMM MMMM Sample sizes for
MMM each group are
SS
SSSS generally proportional
Small SSSS to the relative size of
SS the group

65
Sampling Strategy :
Process Study
Systematic Sampling
Process Sample Description

X X X X X X X X X X X – Sample
every nth
one (e.g.,
4th one)
Must select sampling frequency

66
Sampling Strategy :
Process Sampling
Rational Subgrouping
Process
Hour 1 Hour 2 Hour 3
Sample
Subgroup of
X X X X X X X X X X X X X X X X X X X X X X X X samples

Description
Sample at point “A” in the process every Xth hour
67
Thank you

HAL Management Academy 68

You might also like