Abstract
Robust design methodology (RDM) comprises systematic efforts to achieve insensitivity of products or processes to sources of unwanted variation. In this article, the literature is reviewed and practices that facilitate industrial use of RDM by providing concrete ideas to generate robust designs are identified. To date the literature has focused mainly on statistical techniques useful for creating robust designs, that is, solutions that are insensitive to sources of unwanted variation, while scope and overall framework have been less emphasised, causing an ambiguity in these respects. One practice identified for insensitivity to variation sources is to exploit non-linearities (between response and control factors) and interactions (between noise and control factors), and suitable tools for accomplishing this can be design of experiments or simulation techniques. As systematic RDM efforts are based on an awareness of variation and are beneficial in all design stages, the review also focuses on these two aspects of RDM.
Acknowledgements
We would like to acknowledge the financial support of AB SKF – the knowledge engineering company, the Swedish Foundation for Strategic Research through GMMC, the Gothenburg Mathematical Modelling Center, and VINNOVA – the Swedish Governmental Agency for Innovation Systems – through VMLPD, a project aiming to develop tools for lean product development. We are also grateful to two anonymous referees whose suggestions and comments helped to improve the paper Comments on an earlier draft of this paper made by Bo Bergman also helped to improve it.