Integrating Principal Component Analysis, Fuzzy Linguistic Reasoning and Taguchi Philosophy for Quality-Productivity Optimization

Abstract

The present work demonstrates a multi-response optimization problem for selection of optimal cutting parameters (optimal process environment) for machining (turning) of nylon 6; by using Principal Component Analysis (PCA) followed by fuzzy linguistic reasoning in combination with Taguchi’s robust design technique. In this study, three controllable process parameters: cutting speed, feed, and depth of cut have been considered for obtaining desired Material Removal Rate (MRR) of the process and favorable multiple surface roughness features for the machined product; based on L9 orthogonal array experimental design. The study has been aimed to search an appropriate process environment for simultaneous optimization of quality-productivity favorably. Various surface roughness parameters of statistical importance (of the machined product) have been considered as product quality characteristics whereas; MRR has been treated as productivity measure for the said machining process. To avoid assumptions, limitations, uncertainties and imprecision in application of existing multi-response optimization techniques; Principal Component Analysis (PCA) has been proposed to convert correlated responses into uncorrelated quality indices (called individual principal components); next, a fuzzy inference system (FIS) has been proposed for meaningful and feasible aggregation of individual principal components into an equivalent single quality index, thereby, converting such a multi-objective optimization problem into an equivalent single objective optimization situation. A Multi-Performance Characteristic Index (MPCI) has been defined based on the FIS output. MPCI has been optimized finally using Taguchi method. The study exhibits application feasibility of the proposed approach with satisfactory result of confirmatory test.