A statistics-based fuzzy inference technique to evaluate machining quality level of product involving multiple independent characteristics

Abstract

Manufacturers aim to produce products that meet customer expectations. To this end, they must monitor and assess the machining quality of product to ensure high-quality reliable products. With the rise of consumer awareness, the products that customers can choose from are not only becoming more and more diverse, but also have multiple independent quality characteristics. Therefore, this study adopts the loss-based Six Sigma quality index $Q_{pml}$ , which fully reflects quality level, loss, and yield for machining process of each characteristic l, to develop an assessment model for the machining quality of products with multiple independent quality characteristics of symmetric bilateral specifications. We also construct an analysis chart to simultaneously evaluate and determine the machining performance of all quality characteristics. In practice, $Q_{pml}$ must be estimated based on sample data expressed in precise values. However, uncertainty in sample data is natural and unavoidable in measurement, which increases the risk of misjudgment in assessing machining quality. In view of this, this study proposes the fuzzy statistical testing of $Q_{pml}$ based on the defined fuzzy estimation of $Q_{pml}$ to more accurately determine quality level of machining and reduce consumer risk for each characteristic. A real-world case involving five-way pipe product illustrates the applicability of the proposed technique.