Transcript Stevenppt
Chapter 5 – Part 3
The TQM Philosophy
Mini Case: Quality Improvement
Operation: Adding Toner to Cartridge
Current Process
USL
LSL
20% Defective
Mean
Target
Target Toner
X = Amount of Toner Toner
Mini Case: Quality Improvement
What’s wrong with this operation?
How should it be corrected?
Why is this fix feasible?
Mini Case: Quality Improvement
New Process – Mean Shifted to left
and centered on target
USL
LSL
Target
Amount of Toner
Mini Case: Quality Improvement
Benefits?
Next step?
Mini Case: Quality Improvement
Suppose the current process looked like this.
Will adjusting the mean to the target improve
the process?
LSL
USL
20% Defective
Amount of Toner
Mean
Target
Mini Case: Quality Improvement
Mean adjusted to target
LSL
USL
10%
Defective
10%
Defective
Amount of Toner
Mean =Target
Seven Tools of Quality Control
Cause-and-Effect Diagrams
Flowcharts
Checklists
Control Charts
Scatter Diagrams
Pareto Analysis
Histograms
Cause-and-Effect Diagram
(Fishbone Diagram)
Methods
Materials
Cause
Cause
Cause
Cause
Cause
Cause
Environment
Cause
Cause
Cause
Cause
Manpower
Cause
Cause
Machines
4M + E
Effectproblem
Flowcharts
Checklist
Simple data check-off sheet designed to identify type of
quality problems at each work station; per shift, per
machine, per operator
Control Charts (Chapter 6)
Control
charts are tools for predicting the
future performance of a process.
If we can predicting performance, we
can take corrective action before too
many nonconforming units are produced.
Control Charts (Chapter 6)
Suppose
we construct a control chart for
the thickness of the gold plating of an
electrical connector.
We take samples of connectors over
time and compute the mean of each
sample.
After several time period, we use the
sample means to estimate the mean
thickness.
Control Charts (Chapter 6)
We then construct two control limits:
an upper control limit (UCL) and
a lower control limit (LCL)
• We do this by adding subtracting 3 standard
deviations to the estimated mean:
LCL = Estimated Mean – 3(Standard Deviation)
UCL =Estimated Mean + 3(Standard Deviation)
Control Charts (Chapter 6)
We plot the estimated mean and the control
limits on the control chart.
The result is called a control chart for the
process mean.
Mean thickness
mean
Time
Control Charts (Chapter 6)
If the sample means fall randomly within the
control limits, the process mean is in control.
“In control” means that the process mean is
stable and hence predictable.
If at least one sample mean fall outside of the
control limits, we say the process mean is “out
of control.”
In this case, the process mean is unstable
and not predictable.
The goal is to find out why and remove the
causes of instability from the process.
Scatter Diagrams
A graph that shows how two variables are
related to one another
Speed vs. Yield
30
Yield
25
20
15
10
5
0
0
10
20
Speed
30
40
Optimal Speed
Pareto Diagram
80% of the
problems
may be
attributed to
20% of the
causes.
Percent of defects
Pareto
Principle:
80%
Runs
Bubbles Missing Cracks
Uneven
Histograms
Histogram for Diameter
45
40
35
30
LSL
USL
25
20
15
10
5
0
<=0.077
.077.277
.277.477
.477.677
.677.877
.8771.077
1.0771.277
Diameter
1.2771.477
1.4771.677
1.6771.877
1.8772.077
>2.077
Reliability
Reliability is the probability that the
product, service or part will function
as expected.
Reliability is a probability function
dependent on sub-parts or
components.
Reliability
Reliability of a system is the product of
component reliabilities:
RS = (R1) (R2) (R3) . . . (Rn)
RS = reliability of the product or system
R1 = reliability of the first component
R2 = reliability of the second component
.
. .
Rn = reliability of the nth component
Example 1: Components in Series
A radio has three transistors.
All transistors must work in order for the radio
to work properly.
Probability that the first transistor will work =.80
Probability that the first transistor will work =.90
Probability that the first transistor will work =.85
What is the reliability of the radio?
Solution
R1 = .80
R2 = .90
RS = (R1) (R2) (R3)
RS = (.80) (.90) (.85) =.51
R3 = .85
Example 2: Backup Components
Backup component takes over when a
component fails.
Suppose only one transistor is needed for the
radio to work.
In case the one transistor fails, a backup
transistor has been installed.
Probability that the original transistor will work
=.92
Probability that the backup transistor will work
=.87
Example 2: Backup Components
The backup transistor is in parallel to the
original transistor.
R1 = .92
RBU = .87
Example 2: Backup Components
Parallel components allow system to operate if one
or the other fails
Increase reliability by placing components in parallel
For system with one component and a BU
component:
RS = R1 + [(RBU) x (1 - R1)]
1 - R1 = Probability of needing BU component
= Probability that 1st component fails
Solution
RS = R1 + [(RBU) x (1 - R1)]
R1 = .92
RS = .92 + [(.87) x (1 - .92)]
= .92 + [(.87) x (.08)]
= .9896
RBU = .87
Example 3:
Series with Backup Components
R1 = .80
RBU = .75
R2 = .88
Example 3:
Series with Backup Components
• BU is in parallel to first component.
• Convert to system in series.
• To this by first finding reliability (probability) of
components.
A = Probability that first component or its BU
works
B = Probability that second component works = R2
RS = A x B
Solution
A = R1 + [(RBU) x (1 - R1)]
= .80 + [(.75) x (1 - .80)]
= .95
Part 1
Part 2
.95
.88
B = R2 =.88
RS = A x B = .95 x .88 = .836
Reliability Over Time - Bathtub Curve
Infant Mortality
Failure Rate
Maturity
Constant Failure
t0
t1
t2
Time
Quality Awards and Standards
Malcolm Baldrige National Quality
Award (MBNQA)
The Deming Prize
ISO 9000 Certification
MBNQA- What Is It?
Award named after the former Secretary
of Commerce – Regan Administration
Intended to reward and stimulate quality
initiatives
Given to no more that two companies in
each of three categories;
manufacturing, service, and small
business
Past winners:
Motorola Corp., Xerox, FedEx, 3M, IBM,
Ritz-Carlton
Baldrige Criteria
Leadership (125 points)
Strategic Planning (85 points)
Customer and Market Focus (85 points)
Information and Analysis (85 points)
Human Resource Focus (85 points)
Process Management (85 points)
Business Results (450 points)
The Deming Prize
Given by the Union of Japanese Scientists
and Engineers since 1951
Named after W. Edwards Deming who
worked to improve Japanese quality after
WW II
Not open to foreign companies until 1984
Florida P & L was first US company winner
Based on how well a company applies
Deming’s 14 points
ISO 9000
Set of international standards on quality
management and quality assurance,
critical to international business
Data based approach to decision making
Supplier relationships
Continuous improvement
Customer focus
Leadership
Employee training
Process (operations) management