Desarrollo de un sistema automático para la producción de fibras a partir de botellas plásticas pet

Abstract

Environmental and health pollution is constantly increasing due to its use in different production sectors, which implies its management complexity due to the diversity of plastic products such as PET. In this regard, several studies on the design of plastic shredding machines were found to lack process automation. Thus, this project aims to reuse plastic bottles for fiber extraction through automated mechanical processes such as bottle ring and base clamping, cutting, and slicing. Motors and electronic circuits were adapted to reduce human intervention in the reuse process. A YOLOv5 machine learning (ML) model was also used for object recognition and computer vision to identify bottle dimensions. This allows a user to control the processes of the fiber extraction system. During experimental testing, the AI algorithm showed an accuracy of 91%, a reliable indicator for continuing with subsequent processes, which achieves 80% efficiency, an open field for improving mechanical and electronic processes.