Track 2 Tutorial 2
3:30 to 5:00 p.m. Wednesday June 10, 2009

DOE for Reliability Prediction and Product Performance Optimization

Design of Experiments (DOE) is an indispensable tool to solve complex problems and proactively develop optimal product designs. Multiple product performance requirements (including reliability requirements) must be jointly satisfied, and it is rarely obvious how to optimally meet multiple requirements. DOE, when properly applied, provides an extremely efficient approach to developing product/process understanding using predictive models and allows the identification of optimal solutions (product design specifications and/or process specifications). Unfortunately, when DOE is utilized, it is often misapplied. This tutorial will present the most common misapplications and misunderstandings of DOE. A case study related to automotive window systems will be presented to illustrate successful application of DOE to optimize a manufacturing process in order to achieve reliability and other product performance requirements.

Key Words: Design of Experiments, DOE, Product Optimization, Process Optimization, Multi-Response Optimization

Allise Wachs
Integral Concepts, Inc.
West Bloomfield, Michigan