Orthogonal Robust Design Method for Product Quality

Abstract

The robust parameter design is effective method that is to produce high quality product at low cost to the manufacturer. This paper analyses disadvantages of Taguchi quality method, of which the amount of calculation is increased, algorithm is complicated, and the optimal value of variables in parameter design is updated in regression analysis. Having discovered a lot of useful information not to be utilized in Taguchi quality method, this paper introduces a new robust design method based on orthogonal optimization and variance ratio analysis. The optimal values of parameter and tolerance, at one time, can be obtained by means of successive-orthogonal arrays and variance ratio analysis in parameter design. The best solution in last orthogonal arrays and its tolerance selected by variance ratio analysis make up of factors in successive-orthogonal arrays. The bigger is variance ratio in last orthogonal arrays, the smaller is forecast range of tolerance in successive-orthogonal arrays. It is not necessary to make tolerance design after parameter design and to use crossed array design, signal-to-noise ratio analysis and regression analysis. Numerical result shows that the new method is far simpler and more convenient than Taguchi quality method.