Process improvement.pdf

5. Process Improvement

1. Introduction

Definition of experimental design

2. Assumptions

Measurement system capable

Uses

Process stable

Steps

Simple model

Residuals well-behaved

3. Choosing an Experimental Design

Set objectives

4. Analysis of DOE Data

DOE analysis steps

Select process variables and levels

Plotting DOE data

Select experimental design

Completely randomized

designs

Modeling DOE data

Testing and revising DOE models

Interpreting DOE results

Randomized block designs

Confirming DOE results

Full factorial designs

DOE examples

Full factorial example

Fractional factorial designs

Plackett-Burman designs

Fractional factorial example

Response surface designs

Response surface example

Adding center point runs

Improving fractional design

resolution

Three-level full factorial

designs

10.

Three-level, mixed-level and

fractional factorial designs

http://www.itl.nist.gov/div898/handbook/pri/pri.htm (1 of 2) [5/7/2002 4:01:41 PM]

5. Process Improvement

5. Advanced Topics

When classical designs don't work

6. Case Studies

Eddy current probe sensitivity study

Computer-aided designs

D-Optimal designs

Sonoluminescent light intensity

study

Repairing a design

Optimizing a process

Single response case

Multiple response case

Mixture designs

Mixture screening designs

Simplex-lattice designs

Simplex-centroid designs

Constrained mixture designs

Treating mixture and process

variables

together

Nested variation

Taguchi designs

John's 3/4 fractional factorial

designs

Small composite designs

An EDA approach to experiment

design

7. A Glossary of DOE Terminology

8. References

Click here for a detailed table of contents

http://www.itl.nist.gov/div898/handbook/pri/pri.htm (2 of 2) [5/7/2002 4:01:41 PM]

5. Process Improvement

5. Process Improvement - Detailed Table of

Contents [5.]

Introduction [5.1.]

What is experimental design? [5.1.1.]

What are the uses of DOE? [5.1.2.]

What are the steps of DOE? [5.1.3.]

Assumptions [5.2.]

Is the measurement system capable? [5.2.1.]

Is the process stable? [5.2.2.]

Is there a simple model? [5.2.3.]

Are the model residuals well-behaved? [5.2.4.]

Choosing an experimental design [5.3.]

What are the objectives? [5.3.1.]

How do you select and scale the process variables? [5.3.2.]

How do you select an experimental design? [5.3.3.]

Completely randomized designs [5.3.3.1.]

Randomized block designs [5.3.3.2.]

Latin square and related designs [5.3.3.2.1.]

Graeco-Latin square designs [5.3.3.2.2.]

Hyper-Graeco-Latin square designs [5.3.3.2.3.]

Full factorial designs [5.3.3.3.]

Two-level full factorial designs [5.3.3.3.1.]

Full factorial example [5.3.3.3.2.]

Blocking of full factorial designs [5.3.3.3.3.]

http://www.itl.nist.gov/div898/handbook/pri/pri_d.htm (1 of 6) [5/7/2002 4:01:29 PM]

5. Process Improvement

Fractional factorial designs [5.3.3.4.]

A 2 3-1 design (half of a 2 3 ) [5.3.3.4.1.]

Constructing the 2 3-1 half-fraction design [5.3.3.4.2.]

Confounding (also called aliasing) [5.3.3.4.3.]

Fractional factorial design specifications and design

resolution [5.3.3.4.4.]

Use of fractional factorial designs [5.3.3.4.5.]

Screening designs [5.3.3.4.6.]

Summary tables of useful fractional factorial designs [5.3.3.4.7.]

Plackett-Burman designs [5.3.3.5.]

Response surface designs [5.3.3.6.]

Central Composite Designs (CCD) [5.3.3.6.1.]

Box-Behnken designs [5.3.3.6.2.]

Comparisons of response surface designs [5.3.3.6.3.]

Blocking a response surface design [5.3.3.6.4.]

Adding centerpoints [5.3.3.7.]

Improving fractional factorial design resolution [5.3.3.8.]

Mirror-Image foldover designs [5.3.3.8.1.]

Alternative foldover designs [5.3.3.8.2.]

Three-level full factorial designs [5.3.3.9.]

10.

Three-level, mixed-level and fractional factorial designs [5.3.3.10.]

Analysis of DOE data [5.4.]

What are the steps in a DOE analysis? [5.4.1.]

How to "look" at DOE data [5.4.2.]

How to model DOE data [5.4.3.]

How to test and revise DOE models [5.4.4.]

How to interpret DOE results [5.4.5.]

How to confirm DOE results (confirmatory runs) [5.4.6.]

Examples of DOE's [5.4.7.]

Full factorial example [5.4.7.1.]

Fractional factorial example [5.4.7.2.]

Response surface model example [5.4.7.3.]

http://www.itl.nist.gov/div898/handbook/pri/pri_d.htm (2 of 6) [5/7/2002 4:01:29 PM]

5. Process Improvement

Advanced topics [5.5.]

What if classical designs don't work? [5.5.1.]

What is a computer-aided design? [5.5.2.]

D-Optimal designs [5.5.2.1.]

Repairing a design [5.5.2.2.]

How do you optimize a process? [5.5.3.]

Single response case [5.5.3.1.]

Single response: Path of steepest ascent [5.5.3.1.1.]

Single response: Confidence region for search path [5.5.3.1.2.]

Single response: Choosing the step length [5.5.3.1.3.]

Single response: Optimization when there is adequate quadratic

fit [5.5.3.1.4.]

Single response: Effect of sampling error on optimal

solution [5.5.3.1.5.]

Single response: Optimization subject to experimental region

constraints [5.5.3.1.6.]

Multiple response case [5.5.3.2.]

Multiple response: Path of steepest ascent [5.5.3.2.1.]

Multiple response: The desirability approach [5.5.3.2.2.]

Multiple response: The mathematical programming

approach [5.5.3.2.3.]

What is a mixture design? [5.5.4.]

Mixture screening designs [5.5.4.1.]

Simplex-lattice designs [5.5.4.2.]

Simplex-centroid designs [5.5.4.3.]

Constrained mixture designs [5.5.4.4.]

Treating mixture and process variables together [5.5.4.5.]

How can I account for nested variation (restricted randomization)? [5.5.5.]

What are Taguchi designs? [5.5.6.]

What are John's 3/4 fractional factorial designs? [5.5.7.]

What are small composite designs? [5.5.8.]

An EDA approach to experimental design [5.5.9.]

Ordered data plot [5.5.9.1.]

Dex scatter plot [5.5.9.2.]

http://www.itl.nist.gov/div898/handbook/pri/pri_d.htm (3 of 6) [5/7/2002 4:01:29 PM]

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: