Spring 2008
EMSE 182, 282,
Professor Jim Harris, Control Chart Basics - 1
CONTROL CHART BASICS
PRELIMINARIES
Basic Vocabulary
• Statistical Process Control (SPC)
collection of problem solving tools useful in achieving process
Ä
stability and improving capability through reduced variation.
tools include (The Magnificent Seven)
Ä
1. Histogram
2. Check Sheets
3. Pareto Diagram
4. Cause and Effect Diagram
5. Defect Concentration Diagram
6. Scatter Diagram
7. Control Chart
• A Process "In Control"
one that has a stable mean and variance
Ä
one that operates with only chance variation
Ä
mathematically represented by
Ä
X
,
where {
} are iid with
~ N(0,1)
t
t
t
t
œ
.
5%
%
%

This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*
Spring 2008
EMSE 182, 282,
Professor Jim Harris, Control Chart Basics - 2
• Assignable Cause
a source of variation external to the process caused by wearing
Ä
machinery, operator error, defective raw materials, etc.
usually larger than the chance variation
Ä
a process operating in the presence of one or more assignable causes
Ä
is said to be "out of control"
usually exhibits itself via
Ä
1. sudden or prolonged shift in mean
2. sudden or prolonged shift in variance
3. sudden or prolonged shift in correlation among observations
example models of assignable cause at t
Ä
‡
X
(shift in mean)
t
t
t
t
t
t
t
*
œ
œ
.
5%
.
$
5%
‡
X
(shift in variance)
t
t
t
t
t
t
t
*
œ
œ
.
5%
.
$5%
‡
X
(linear trend in mean)
t
t
t
t
t
t
t
t
*
œ
œ
.
5%
.
α
"
5%
‡
X
(correlation in observations)
t
t
(X
t
t
t
t
t
1
t
*
œ
Ñ
œ
.
5%
.
9
.
5%
‡

Spring 2008
EMSE 182, 282,
Professor Jim Harris, Control Chart Basics - 3
STATISTICAL BASIS FOR CONTROL CHARTS
Basic Design
• On Chart
data plotted and connected with line segments
Ä
center line, upper and lower control limits (UCL and LCL), and
Ä
sometimes upper and lower warning limits plotted on chart.
• Use
process is out-of-control, look for assignable cause when
Ä
1. any point plots outside the control limits outside
2. certain runs criteria are observed
3. any pattern in data is observed

This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*