文档介绍：Edited by Foxit PDF Editor
Copyright (c) by Foxit pany, 2004 - 2007
For Evaluation Only.
Design of Experiments Pat Hammett
Introduction to Design of
Experiments (DOE)
1
Experimental Design Variables
Process Input Variables (Factors) Process Output Variables
Control Variables (Design Factors)
- signal factors
(location or mean effects)
Manufacturing Process Output, or
- noise factors Process Response Variable, or
(dispersion or variation effects) Product Characteristic
Uncontrollable Variables
- environmental factors
those that are difficult or very
expensive to control.
(., ambient temperature)
2
1
TQM - University of Michigan
Design of Experiments Pat Hammett
Duality of Signal & Noise Factors
Effect of Input Variables on a Process Output Adjustment Factors
affects the mean of the process
without affecting variation “knob”.
Need these to center a process.
Noise Duality Signal (especially if Pp is high, Ppk low)
Factors location & Factors Example:
(dispersion effects) dispersion (location effects) tool position knob in machining
Dispersion Effect?
Yes No
Location Yes Duality Factor Adjustment Factor
Effect? No Noise Factor Robust Factor
3
Operating Windows
Defined:
Output Variable
Input Variable
4
2
TQM - University of Michigan
Design of Experiments Pat Hammett
Summary of Terminology
n Most input variables have a robust operating
window in which the output is insensitive or
able to meet manufacturing specifications.
(., relatively few pure linear relationships).
Output Variable Input Variable
5
Experimental Design and DFSS
n In Design for Six Sigma (DFSS), use DOE to:
n Identify key input variables which affect the mean or
variation of a key product output variable.
n After determining key inputs, mfgs should either:
n Identify robust levels and establish controls to maintain,
n Fix the settings for an input variable, or
n Remove input variabl