Title: STATISTICAL PROCESS CONTROL AND QUALITY MANAGEMENT
1STATISTICAL PROCESS CONTROL AND QUALITY MANAGEMENT
2Quality
- Fitness for use, acceptable standard
- Based on needs, expectations, perceptions and
customer requests
3Types of Quality
- Quality of design
- Quality of conformance
- Quality of performance
4Quality of Design
- Differences in quality due to design differences,
intentional differences
5Quality of Conformance
- Degree to which product meets or exceeds standards
6Quality of Performance
- Long term consistent functioning of the product,
reliability, safety, serviceability,
maintainability
7Quality and Productivity
- Improved quality leads to lower costs and
increased profits
8Statistics and Quality Management
- Statistical analysis is used to assist with
product design, monitor the production process,
and check quality of the finished product
9Checking Finished Product Quality
- Random samples selected from batches of finished
product can be used to check for product quality
10Assisting With Production Design
- A variety of experimental design techniques are
available for improving the production process
11Monitoring the Process
- Control charts can be used to monitor the process
as it unfolds - Sampling from the production line to see if
variation in product quality is consistent with
expectations
12The Control Chart
- A special type of sequence plot which is used to
monitor a process - Throughout the process measurements are taken and
plotted in a sequence plot - Plot also contains upper and lower control limits
indicating the expected range of the process when
it is behaving properly
13Examples
14Examples
15Examples
16Examples
17Examples
18Two Types of Variability in Control Charts
- Chance or Common Causes
- Special or Assignable Causes
19Chance or Common Causes
- Numerous small causes which are inherent to the
process and which occur at random - Considered to be normal variation when process is
under control
20 Special or Assignable Causes
- Have large effects on the process
- Occur rarely or sporadically
21Statistical Control/Stable Process
- Variability only due to common or chance causes
- Variability natural, expected
- Variability managed
22Out of Control Process
- When special causes or assignable causes are
present
23Types of Control Charts
- Control Chart For The Mean
- Control Chart For The Range
- Control Chart For A Proportion
24Control Chart For The Sample Mean
- Assuming that the sample mean is approximately
normal with mean ? and standard deviation ?, a
control chart for the mean usually consists of
three horizontal lines - The vertical axis is used to plot the magnitude
of observed sample means while the horizontal
axis represents time or the order of the sequence
of observed means
25Control Chart For The Sample Mean
- The center line is at the mean, ? and upper and
lower control limits are at - 3 and -3
- Since ? is usually unknown the
- term is usually replaced by an
estimator based on the sample range (see below)
26Control Chart For The Sample Mean
- The formula is given by
- where average value of the range
-
- k number of samples
- The values of A2 depend on the sample size
27Control Chart For The Sample Mean
28Control Chart For The Range
- Designed to monitor variability in the product
- Range easier to determine than standard deviation
- Distribution of sample range assumes product
measurement is normally distributed
29Control Chart For The Range
- Upper and lower control limits and centre line
obtained from and the values of D3, D4 - according to the formulae
- LCL
- UCL
- The values of D3, D4 are based on sample size and
are given in a table below
30Examples
31Examples
32Control Chart For The Sample Proportion
- Population proportion ?
- The sample mean is now a mean proportion given by
- Where total number of objects in sample
with characteristic divided by total sample size
33Control Chart For The Sample Proportion
- Using the central limit theorem the control
limits are given by
34Control Chart For The Sample Proportion
- Since the true proportion ? is usually unknown we
replace it by the average proportion - If the sample size varies then the upper and
lower limits will vary and the equations become - where ni sample size in sample i
35Examples