Pattern recognition in classification of intelligent composites during smart manufacturing

Abstract

The development of an on line computer based classification system for the real time classification of different composites is addressed in this study. Different parameters were collected simultaneously when embedded sensors (dielectric, optical fibre, and piezoelectric sensors) were used within two different composite matrices during the curing process. The measurements were used by an algorithm-based software as a logged data file, resulting in to induction of a decision tree. Later, a systematic software is designed based on the rules derived from this decision tree, to recognise the type of composites used in the experiment together with recognition of their physical and mechanical characteristics. This is a new approach to data acquisition in intelligent materials produced by smart manufacturing system. Keywords. Computer modelling, process controls, smart composites, smart manufacturing. © Springer-Verlag 2003.