Automating reduction cell thermal monitoring

Abstract

Kerlekar eru vel þekkt vandamál við álframleiðslu um allan heim, Þegar ker leki verður skapast töluverð hætta fyrir starfsmenn ásamt því að hlotist getur mikið fjarhagslegt tjón af lekanum sem gerir það að verkum að mikil áhugi er til staðar fyrir lausnum sem koma í veg fyrir að kerleki verði.

For decades, industrial robots have been widely used in factories where repetitive actions and tasks are automated, replacing manual labor. With increasing automation and cheaper and more capable technologies, the manufacturing industry has become more reliant on intelligent industrial robots. This project is a BSc thesis in applied electrical engineering at Reykjavík University and is performed in cooperation with Alcoa Fjarðaál. The motivation for this project comes from the author’s experience working with the emergency maintenance crew at Alcoa and is inspired by Guðjón Hugberg Björnsson’s master’s thesis, “Automatic Thermal Inspection of Aluminum Reduction Cell,” [1] which analyzes thermal images of reduction cells by using modern computer vision techniques to aid in cell leak detection.