Search

This paper regulates a rational approach of robust design in material machining process on basis of the probability method for multi-objective optimization. As an expansion, it takes the arithmetic mean value of the performance index of the alternative and its deviation as twin independent responses to make contributions of two parts of the partial favorable probability individually; the arithmetic mean value of the performance index represents the performance index according to its function and preference in the assessment, and the deviation exhibits an index with the characteristic of the smaller the better in general. Furthermore, the product of the above twin parts of partial favorable probability forms the actual favorable probability of the performance index of each alternative in the treatment, which is the unique indicator of the performance index of the alternative in the robust optimum. Moreover, examples are given to illustrate the application of the approach. [ABSTRACT FROM AUTHOR]

A novel approach for material selection is developed on basis of probability theory. The novel concept of favorable probability is formed to express the favorable degree of a candidate material in the selection. It assumes that all utility indexes of material performance indicators can be attributed to beneficial or unbeneficial type, and each utility index of material performance indicator devotes to one partial favorable probability in the selection; a partial favorable probability is linearly correlative to the corresponding utility index of material performance indicator in positive or negative manners upon its nature of whether beneficial or unbeneficial type; the product of all the partial favorable probabilities forge the overall / global favorable probability of a candidate material. Finally, all the candidate materials can be ranked according to their overall favorable probability comparatively in the material selection. As application examples, the quantitative material selections of cylindrical shaft, stamping of titanium bipolar plates for fuel cell and tribological coatings are given; furthermore, the novel approach is combined with the orthogonal and uniform experiment designs to conduct their optimizations under multi‐object conditions. [ABSTRACT FROM AUTHOR]