SPC Tutorial 1

You might also like

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 15

An Introduction To SPC

What is SPC?

• Statistical
• Process
• Control
What are Statistics

• The science of collecting, organizing, and analyzing data.


• Numbers or values that help to describe the characteristics of a
selected group; technically, statistics describe a sample of a
population.
Populations and Samples
Population:
•The entire set of data from
which a sample is selected
and about which we wish to
draw conclusions.

Sample:
If the cake is the POPULATION…
•A part of a population
taken to be representative
of the population as a
whole for the purposes of
investigation.

From a test of the sample,


we could deduce
That the whole of the cake A slice would be the SAMPLE.
contained chocolate!
Sampling Methods
Simple Sampling Stratified Sampling
All elements of population Elements are chosen
are treated equally when proportional to appearance in
selecting. population.

Random Sampling Example.


When choosing a sample of
Required number of elements are selected entirely at random from
20 people from a
the population or stratified subgroup population
containing 70% women, we
Systematic Sampling might select 14 women and 6
men.
Sample elements are taken at regular intervals from the population
based on a random start point.

Convenience Sampling
Samples are chosen at intervals convenient to the sampler.
Sampling at Fibrax
• Population is Stratified (Tool Cavity)
• Systematic Sampling is used based on time (hourly, two hourly etc.)
What can we do with our Sample?
• Sample represents Population.
Population.
• Facts determined about
about Sample
Sample cancan be
be used
used toto assess
assess
Population.
• Initial things to look for are the
the ‘Distribution’
‘Distribution’ of
of the
the data
data and
and the
the
resultant ‘Descriptive Statistics’.
Statistics’.
Distribution
• The process (and resultant parts) will vary from cycle to cycle, this is known as
Common Cause Variation.
• Because of this variation, no two parts will be identical but will conform to a
Distribution Pattern.
• Most processes conform to what is known as a Normal Distribution.
The Normal Distribution
• Conforms to ‘Bell’ Shaped Curve.
• Predictable
• Can be described by just two statistics.
– Mean
– Standard Deviation.
Descriptive Statistics
• Central Tendency
– Mean: The sum value of the sample divided by the number of elements in it.
• E.g. Numbers 1,2,3,7,4,3,9,8: Mean is (1+2+3+7+4+3+9+8)/8 = 4.625
– Mode: The most frequently occurring value in the sample.
• E.g. Numbers 1,2,3,7,4,3,9,8 : Mode is 3
– Median: The value which has half the scores placed above, and half below it.
• E.g. Numbers 1,2,3,7,4,3,9,8 : Median is 1,2,3,3 --- 4,7,8,9 = 3.5
• Spread
– Range: The Maximum value in the sample minus the minimum value.
– Standard Deviation: An index of the average displacement of each value from the
mean.
• Skew
– Whether more values appear to one end or other of the range of values.
• Kurtosis
– Whether the distribution is truncated (missing an end).
Descriptive Statistics

99.98%
95.4%
68.2%

Mean

1 1
STANDARD STANDARD
DEVIATION DEVIATION

2 2
STANDARD STANDARD
DEVIATIONS DEVIATIONS

3 3
STANDARD STANDARD
DEVIATIONS DEVIATIONS
So What?
• If the process fails to conform to this pattern, this indicates the
presence of Special Cause Variation.
• Special Cause Variation indicates that something in the process
has changed. This may be beneficial to the process but usually
isn’t.
Control Charts
• Graphical representation of process distribution
• Includes representation of time
• Various types of Chart available
• Xbar (Mean) and Range most common
• Allow us to monitor the process to look for special causes.

•Measure Sub-group
of parts (usually 5)
•Calculate the Mean
(Xbar)
•Plot on top Chart

•Calculate Range (R)


•Plot on bottom chart

•Check for Rule breaking


•Divide Chart into 3 Zones
The Rules
•No point should be outside Zone C
(More than 3 Standard Deviations from Mean)

•No more than 2 points out of 3 should be


outside Zone B on one side
(More than 2 Standard Deviations from Mean)

•No more than 4 points out of 5 should be ZONE


outside Zone A on one side ZONE C
A
B
(More than 1 Standard Deviations from Mean)

•No more than 7 points in a row outside


Zone A
(More than 1 Standard Deviations from Mean)

•No more than 8 points in a row on one


side of centre line (mean)
•No more than 5 points continually ZONE C
increasing or decreasing ZONE B
A
•No more than 14 points in a row
alternating up and down
•No more than 14 points in a row in Zone A
Summary
• Statistics allow measurement of a small sample to predict the
behaviour of a larger population.
• Every process is affected by Common Cause Variation.
• Most processes conform to the Normal Distribution.
• Failure to conform to the Normal Distribution is the result of
Special Cause Variation.
• The behaviour of the process can be monitored using Control
Charts.

You might also like