No Cover Image

Conference Paper/Proceeding/Abstract 637 views 108 downloads

Flame intensity analysis for hot molten metal pouring in the steel industry by applying image segmentation

Ivan Popov, Grazia Todeschini

Global Congress on Manufacturing and Management, 20-22 April 2021

Swansea University Author: Grazia Todeschini

Abstract

Pouring large quantities of hot metal (HM) can release substantial amounts of flame. This problem is frequently encountered within the Basis Oxygen Furnace (BOF) steelmaking process where large quantities of HM (frequently exceeding 300 t) are poured into the converter vessels. The HM is contained i...

Full description

Published in: Global Congress on Manufacturing and Management, 20-22 April 2021
Published: 2021
URI: https://cronfa.swan.ac.uk/Record/cronfa55771
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: Pouring large quantities of hot metal (HM) can release substantial amounts of flame. This problem is frequently encountered within the Basis Oxygen Furnace (BOF) steelmaking process where large quantities of HM (frequently exceeding 300 t) are poured into the converter vessels. The HM is contained in specially designed ladles and poured using overhead girder cranes. Excess release of flame may damage surrounding components such as crane ropes and consequently reduce their lifecycle. Therefore, limiting the release of flame during pouring, allows extending the lifetime of the components located in proximity of the ladle. The scope of this paper is to characterise flame generation during different pouring operations at a BOF steelmaking plant and to relate the amount of flame generated to process factors. Due to the complexity of the process under investigation, this paper does not aim to eliminate flame generation, but rather to identify approaches to its mitigation. The proposed approach utilises a standard CCTV camera to record videos of pours. An image segmentation analysis is then performed, where the flame is separated from the background image using pixel information in the CIE L*a*b* colour space. For each frame, flame intensity is then calculated. This process is partially automated for each video making use of MATLAB. A total of 169 videos are analysed and the pours that cause higher flame intensity are identified. In the last steps of the analysis, the process factors with the most significant impact on the flame release are identified and mitigating solutions are proposed.
College: Faculty of Science and Engineering

Similar Items