A mathematical model for simultaneously evaluating design-set options for yield, reliability and compatibility requirements

• Main Author: Lao, Mark Ian Y.
• Format: text
• Language: English
• Published: Animo Repository 2010
• Subjects: Product design; Product development; Product design. and others; Operations Research, Systems Engineering and Industrial Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_masteral/3848; https://animorepository.dlsu.edu.ph/context/etd_masteral/article/10686/viewcontent/CDTG004709_P.pdf
• Institution: De La Salle University

Majority of the new products introduced today come from Reintegration and Limited Innovation (RLI) product development projects. These refine and reuse mature and well known product architectures and design solution and is the focus of the study. The goal of product development projects is to reach full scale production in the least possible time with the best product and process design. During development, numerous conflicting product and process design requirements must be managed in order for the product to be successful. First pass yield, reliability and compatibility are the requirements tackled in this study. Ideally, a design which is high reliability, high yield and high compatibility is selected but the presence of variation hinders this and as such, design iterations which cause delays and cost overruns are common in development. A mathematical model for simultaneously evaluating design-set options for yield, reliability and compatibility requirements was formulated with aiding the development process of RLI projects in mind. The objective of the model is to maximize the end product yield of selected product and process designs. Product reliability and assembly process compatibility are constraints in the model. The major decision variables include subcomponent selection, process resource selection and material selection. iv The general mixed integer non-linear model initially formulated was found to be non-convex and was linearized to aid solution generation accuracy and reliability. A three-process, five-subcomponent product was used to validate the model and analyze its sensitivity. The model was run under different compatibility and reliability requirements. The results were compared with a complete enumeration of the solutions to ensure that optimal solutions were being generated. Design selection patterns and the yield, reliability and compatibility tradeoffs which occurred were analyzed primarily in the results of the runs. Findings from the sensitivity analysis reiterate the importance of exploring all possible options of a given product and process design during development, which goes against the traditional serial method.