Zero-Defect Manufacturing in the Industry 4.0 Era: A Hybrid Approach to Detecting Quality Shifts

Abstract

Thanks to the emergence of the Industry 4.0 paradigm, the idea of Zero-Defect Manufacturing (ZDM) has now been claimed to be possible. The ZDM is based on four strategies: detecting, repairing, predicting, and preventing. The main objective of the present work is to defend the proposition that detecting small changes in product quality is a necessary condition to prevent the generation of non-conformance and consequently implement the zero-defect concept. The traditional statistical control charts - such as the Shewhart, EWMA, and CUSUM - and a new method known as ‘concept drift’ were assessed as candidates to build a Decision Support System (DSS). A prototype of the designed DSS was applied to detect quality changes of tubular steering input shafts produced by end-forming. The combination of a concept drift algorithm with a Shewhart chart presented a better performance when compared to the individual behaviour of the statistical control charts.