Design of tolerances for quality

Design of a product (process) includes system design, parameter design, and tolerance design. Robust design is closely applicable to parameter design and tolerance design. The current literature on robust design is focused on parameter design while little attention has been paid to tolerance design. The current literature on tolerance design has focused on the use of optimization to minimize cost while little attention has been paid to minimizing manufacturing variations. This paper attempts to apply a robust design method, the design of experiments (DOE) approach, in tolerance synthesis to minimize manufacturing variations in the probabilistic case. Both, the manufacturing cost and the number of manufacturing defects are minimized in robust tolerance design. A solution procedure is proposed to apply the DOE approach to probabilistic tolerance design. The procedure is illustrated with an example. Special applications of the DOE approach to tolerance design are discussed. A brief comparison of the DOE approach with optimization, Taguchi methods, and zero defect design is presented.