QUALITY DESIGN METHOD USING PROCESS CAPABILITY INDEX

Abstract

Product quality is usually variable because errors of machining and an assembling are also variable. Tolerancing in a mechanical design process is important because it directly affects product quality and manufacturing cost. Many tolerancing methods have been studied. However, information about conditions and requirements for manufacturing process is rounded to one scalar value as a conventional tolerance. Therefore, authors have proposed statistical tolerance index as a new method, which specify limits of process capability index with traditional tolerance. Efficiency of the method has been confirmed under a certain condition. In this study, design method of statistical tolerance index, which is inverse problem of the stack up problem, is proposed. The proposed method is consisting of Monte Carlo simulation and real encoded genetic algorithm. Efficiency of the proposed method is evaluated through case study.