Soft sensor of bath temperature in an electric arc furnace based on a data-driven Takagi–Sugeno fuzzy model

Blažič, Aljaž (author), Škrjanc, Igor (author), Logar, Vito (author)

Electric arc furnaces (EAFs) are intended for the recycling of steel scrap. One of the more important variables in the recycling process is the tapping temperature of the steel. Due to the nature of the process, continuous measurement of the melt temperature is complicated and requires sophisticated measuring equipment; therefore, for most EAFs, separate temperature samples are taken several times before the melt is tapped, to verify whether the melt temperature is within the prescribed range. The measurements are obtained using disposable probes; when measurement is performed, the furnace must be switched off, leading to increased tap-to-tap time, unnecessary energy losses, and consequently, lower efficiency. The following paper presents a novel approach to EAF bath temperature estimation using a fuzzy model soft sensor obtained using Gustafson–Kessel input data clustering and particle swarm optimization of model parameters. The model uses the first temperature measurement as an initial condition, and measurements of the necessary EAF inputs to estimate continuously the bath temperature throughout the refining stage of the recycling process. The results have shown that the prediction accuracy of the proposed model is very high and that it fulfils the required tolerance band. The model is intended for parallel implementation in the EAF process, with the aim of achieving fewer temperature measurements, shorter tap-to-tap times, and decreased energy losses. Furthermore, if information about bath temperature is accessible in a continuous manner, operators can adjust the control of the EAF to achieve optimal tapping temperature and thus higher EAF efficiency.

• Language: English
• Keywords: electric arc furnace, fuzzy modeling, Gustafson–Kessel clustering, particle swarn optimization, temperature estimation
• Work type: Article
• Typology: 1.01 - Original Scientific Article
• Organization: FE - Faculty of Electrical Engineering
• Publication status: Published
• Publication version: Version of Record
• Year: 2021
• Number of pages: 11 str.
• Numbering: Vol. 113, pt. B, art. 107949
• PID: 20.500.12556/RUL-138854
• UDC: 681.5:621.365.2
• ISSN on article: 1568-4946
• DOI: 10.1016/j.asoc.2021.107949
• COBISS.SI-ID: 79852547
• Publication date in RUL: 23.08.2022

Record is a part of a journal

• Title: Applied soft computing
• Publisher: Elsevier
• ISSN: 1568-4946
• COBISS.SI-ID: 16080679

Licences

• License: CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
• Link: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Description: The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.

Secondary language

• Language: Slovenian
• Keywords: elektroobločna peč, mehko modeliranje, rojenje Gustafson-Kessel, optimizacija z roji delcev, ocena temperature

Projects

• Funder: EC - European Commission
• Funding programme: H2020
• Project number: 869815
• Name: Optimization and performance improving in metal industry by digital technologies
• Acronym: INEVITABLE

• Funder: ARRS - Slovenian Research Agency
• Project number: P2-0219
• Name: Modeliranje, simulacija in vodenje procesov