Petr: automated polyethylene terephthalate (pet) recycler pultrusion machine with a diameter consistency pid control system using computer vision/
Maria Jeriah V. Canta, Jessie F. Logrono, Lieneth A. Maylas, Cherie Ann C. Nelmida, and Vince Elrey O. Tolledo.--
- Manila: Technological University of the Philippines, 2025.
- xi, 168pages: 29cm.
Bachelor's thesis
College Of Engineering.--
Includes bibliographic references and index.
The rapid rise in plastic production poses a serious threat to aquatic ecosystems, with an estimated 13 to 19 million metric tons of plastic entering water bodies each year. Asia alone contributes approximately 82 million tons of plastic waste, making it the largest source of global plastic pollution. As the demand for sustainable materials in 3D printing continues to grow, there is an urgent need to develop methods for converting plastic waste into alternative filament sources. This study addresses that need by designing and building a 3D printer filament pultrusion machine that features automation and real-time monitoring. The system uses PID control through an Arduino Uno R4 microcontroller to regulate the stepper motor, heating nozzle, and spooling mechanism, which all work together to produce consistently wound PET filament. In addition, a computer vision system using machine learning was set up on a Raspberry Pi 5 to monitor filament diameter in real time. The goal was to maintain a consistent 1.75 mm diameter, with an acceptable 5% margin of error (1.66mm to 1.84mm). YOLOv8 was used instead of Mask R-CNN because it worked faster and more efficiently on the Raspberry Pi. To measure the filament, OpenCV was used, offering a more affordable method compared to laser systems. The tests used Nature Spring 1-liter PET bottles, cut into strips with widths between 9.0 mm to 9.6 mm and a thickness of 0.2 mm. It was found that the width and thickness of the strip had an effect on whether the filament could consistently reach the target diameter. Consequently, to check the quality of the filament, a pendulum impact test was done, along with test prints using a 3D printer. Measurements from the computer vision system were compared with results from a caliper, and had an accuracy of 98.73%. This project shows how waste PET bottles can be turned into usable 3D printing filament.
Plastic production Aquatic ecosystems Filament machine