quality improvement introduction to business statistics, 5e kvanli/guynes/pavur (c)2000...
Post on 19-Dec-2015
217 views
TRANSCRIPT
Quality Improvement
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Definitions
• Quality product or service: A product or service that meets or exceeds the expectations of the customer.
• Process: Any combination of people, machinery, material, and methods that is intended to produce a product or service.
• Quality Characteristics: Features of a product that describe its fitness for use .Introduction to Business
Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Definitions
• Statistical Process Control (SPC): The application of statistical quality-
control methods to measure and analyze the variation found in a process.
• Control Chart: A statistical chart used to monitor various aspects of a
process and to determine if the process is in control or out of control.Introduction to Business
Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Malcolm Baldrige National Quality Award Criteria
• Leadership System
• Strategic Planning
• Customer and Market Focus
• Information and Analysis
• Human Resource Focus
• Process Management
• Business ResultsIntroduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Quality Improvement Tools
• Flowcharts
• Cause-and-Effect Diagrams
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Flowchart
Figure 12.2
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Cause-and-Effect Diagram
Figure 12.3
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Variation and Control Charts: Sources of Variation
• Machinery
• People
• Materials
• Production Methods
• The Environment
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Deming Funnel Experiment: Strategies
• Strategy 1: Do not react to this random variation and do not move the funnel.
• Strategy 2: Measure the distance from the marble’s resting place to the bull’s-eye. Move the funnel and equal distance, but
in the opposite direction.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Deming Funnel Experiment: Strategies
• Strategy 3: Measure the distance from the marble’s resting place to the
bull’s-eye. Move the funnel this distance, in the opposite direction, starting at the bull’s- eye.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Deming Funnel Experiment
Figure 12.5
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Control Charts
• A process is in control if the observed variation is due to inherent or natural variation. This variability is the cumulative effect of many small, essentially uncontrollable, causes.
• A process in out of control if a relatively large variation is introduced that can be traced to an assignable cause.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
General Form of a Control Chart
Figure 12.6
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Control Chart
Figure 12.7
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
X and R Charts
X X 1 X 2 X m
m
ˆ R d2
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
X and R Charts
Tables 12.2 & 12.3
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
X and R ChartsProcess for Estimating
• Determine the average of the m values of R.
• Select the values of d2 from Table 12.3 using the corresponding sample size, n.
• Estimate using:
ˆ R d2
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
X and R ChartsControl Limits
UCL X 3ˆ n
X 3( R / d2)
n
Center Line X
LCL X 3ˆ n
X 3( R / d2)
n
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
X and R ChartsControl Limits
Figure 12.9
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
R Chart
sR R d3
d2
UCL R 3sR R 3R d3
d2
1 3
d3
d2
R
LCL R 3sR R 3R d3
d2
1 3
d3
d2
R
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
R Chart
D3 1 3d3
d2and D4 1 3
d3
d2
By defining
UCL D4 R
Center Line R
LCL D3 R Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Steps for Making X and R Charts
• Collect m samples of data, each of size n.
• Compute the average of each subgroup.
• Compute the range for each subgroup.
• Find the overall mean.
• Find the average range.
• Estimate .
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Steps for Making X and R Charts
• Compute the 3-sigma control limits for X.
• Compute the 3-sigma control limits for R
• Construct the control charts by plotting X and R points for each subgroup on the same vertical line.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Pattern Analysis for X
Figure 12.12
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Minitab X Chart
Figure 12.13
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Control Chart for the Proportion Nonconforming: The p chart
Reasons for Using a p Chart
• Quality measurements are not possible.
• Quality measurements are not practical.
• Many characteristics on each part are being judged during inspection
• The main question of interest is: “will the process be able to produce conforming products over time?”
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Steps for Making p Charts
• Collect m samples of data, each of size, n.
• Determine the proportion nonconforming for each sample.
• Find p, the overall proportion nonconforming.
• Compute the 3-sigma control limits
• Draw the control lines and plot the values of pi.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
p Chart Equations
UCL p 3p (1 p )
n
CL p
LCL p 3p (1 p )
n
pi Ti
n
p Ti
mnp
total number of nonconforming units
total sample size
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
p Chart for Ex. 12.3
Figure 12.17
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
The c Chart Construction
• Collect m samples of data, each of size, n.
• Determine the number of nonconformities for the ith unit. Call this value ci.
• Find the average number of nonconformities per unit, c.
• Compute the 3-sigma control limits.
• Construct the chart.Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
c Chart Equations
c ci
m
UCL c 3 c
Center Line c
LCL c 3 c Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Capability
• Specification Limits: process requirements
• Lower spec limit (LSL): the lower limit of the process output that meets the process requirements.
• Upper spec limit (USL): the upper limit of the process output that meets the process requirements.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Capability
Figure 12.20
The difference between 12.01 and 12.19 is referredto as the process spread.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Capability
Figure 12.21
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Capability Ratio Cp
Assumptions:
• The process is centered within specifications.
• The process is normally distributed.
• The process is stable.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Cp
Cp USL LSL
6 ˆ
Cp USL X 3 ˆ
(upper spec limit only)
Cp X LSL
3 ˆ (lower spec limit only)
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Process Capability Ratio Cpk
Assumptions
• The process may or may not be centered in spec.
• The process is normally distributed.
• The process is stable.
• Control charts will be used to monitor the process over time.
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Figure 12.22
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Procedure for Finding Cpk
1. Determine RL X LSL
3 ˆ
2. Determine RU USL X
3ˆ 3. Cpk Minimum of RL and RU
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing
Figure 12.23
Introduction to Business Statistics, 5e
Kvanli/Guynes/Pavur
(c)2000 South-Western College Publishing