The Robust Design Of The Issue Of Uncertainty In Research And Applications

Robust design, a valid design technical, which can assure product quality has been being regarded and applied in electronic industry, engineering industry and so on. However most of achievements have been accomplished under traditional hypothesis or certainty. Since production in practice is usually not accord with those condition. So uncertainty researches need studying. This thesis will research two facets to robust design: experiment design approach and engineer model approach.Four aspects of robust design are researched mainly. Firstly, based on traditional response surface approach the thesis researches response surface under heterogeneity by means of General Linear Model (GLM), moreover establishes one-order response surface model in noise factors and second-order response surface model in noise factors under heterogeneity, and provides model's statistical analysis method. Secondly, based on response surface model in noise factors the thesis researches robust design response surface model with dynamic characteristics, and establishes robust design model with dynamic characteristics, then provides model's statistical analysis method under homogeneity and heterogeneity. The parameter estimation and hypothesis test of the model under heterogeneity are reviewed. Thirdly, the thesis gives a brake system example to illustrate this method, and establishes robust design model with dynamic characteristics and estimate joint mean-dispersion model by use of GLM by SAS. At last the thesis also researches optimal robust design based on cost-quality model via fuzzy mathematics knowledge and gives an engineer model approach. A new optimal robust design method based on experiment design approach and fuzzy mathematics knowledge is brought forward, and the effectiveness of this method is verified by means of a living example.