Deep learning based image segmentation for metal manufacturing
||Deep learning based image segmentation for metal manufacturing|
||NAISS Small Compute|
||Sudhanshu Kuthe <firstname.lastname@example.org>|
||Kungliga Tekniska högskolan|
||2023-05-23 – 2024-06-01|
Electric Arc Furnace (EAF) steelmaking is critical to the European steel industry in the era of the green transition. Optimizing the EAF process is key to achieving high efficiency, low costs, and high-quality steel production. However, the current method of determining the optimal scrap composition for the EAF is very complex and time-consuming due to the involvement of various processing constraints. Understanding these process constraints to optimize scrap melting presents significant challenges. The objective of the proposed project is to design an Artificial Intelligence (AI)-based system that operates on the principle of image segmentation. This system will assist in understanding and optimizing the scrap composition. Moreover, the segmented images are further processed using advanced deep-learning techniques to identify relationships between scrap chemistry and also help in understanding any process constraints observed during scrap melting. The project team will explore optimized functionalities for robust scrap melting technologies useful in EAF steelmaking. Our work also comprises energy-related research, specifically optimizing the electrical energy required for scrap melting. We model, compute, and analyze real-industrial scenarios to understand and support experimentalists and researchers working in this emerging field.