Development of a Rapid Prototyping System using Response Surface Methodology

Abstract

This case study presents an investigation of the relationships between eight process operating variables (factors) and five part performance measures (responses) in a rapid prototyping system. The use of fractional factorial, single-factor foldover, and central composite designs is demonstrated. Polynomial regression models are constructed for each response, followed by a desirability function model. Canonical and ridge analyses are used to identify a group of factor settings that simultaneously produce improved performance for all responses.

An improved dynamic structure-based neural networks determination approaches to simulation optimization problems

“…RSM carried out on the entire domain of interest results in what is called a metamodel [46]. The two most common ways of obtaining this metamodel are regression and neural networks [47,48]. Once a metamodel is in hand, optimization can be carried out using deterministic optimization procedures.…”

An interactive desirability function method to multiresponse optimization

“…It should be noted that these requirements are slightly different from the original ones given in Palmer et al (2006) for an illustrative purpose. The most desirable values of y 1 , y 2 , and y 3 are 0.…”

“…The ''rapid prototyping system" problem is concerned with a new rapid prototyping process, called Selective Inhibition Sintering, that fabricates a part in a layer-by-layer fashion (Palmer et al, 2006). The objective of the problem is to improve the quality of the fabricated part.…”

“…The rapid prototyping process problem has been introduced and solved using the conventional desirability function approach in Palmer et al (2006).…”

“…The (coded) experimental region X is given as: À2 6 x 1 6 2, À2.333 6 x 2 6 2.333, À2.333 6 x 3 6 2.333, À1 6 x 4 6 1, À2.333 6 x 5 6 2.333, À2 6 x 6 6 2, À2.2 6 x 7 6 2.2, À2.5 6 x 8 6 2.5. The estimated response models (Palmer et al, 2006) are given below:…”