Noise Strategy in Robust Design: What Aspects of Noise Factors are Important in Quality Engineering?

Abstract

The correlation and variances of noise factors induced during robust design may influence the success of the quality engineering process. A model-based approach is employed to quantify these influences. The results are then checked using two case studies. It was determined that for systems having no significant three-factor interactions, correlation among the noise factors can be safely neglected and it is often helpful to amplify the magnitude of induced noise. Otherwise, noise factors should be used that match the true magnitude and correlation in field conditions. Based on these results, a procedure is suggested for a noise strategy in quality engineering.

Keywords:

• Hierarchical probability model
• Latin hypercube sampling
• Noise factor surrogates
• Robust design

ACKNOWLEDGMENTS

The support of the Ford-MIT Alliance is gratefully acknowledged.