BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
BEGIN:VEVENT
ORGANIZER;CN=ESTAD 2023:mailto:info@metec-estad.com
LOCATION:Room 17b
SUMMARY:Deep learning based approach to improve the foamy slag quality by optimizing additive injection rate for DRI melting process 
DESCRIPTION:In Electrical Arc Furnace (EAF) melting process, foamy slag quality has a key role in green steel production. A good quality of foamy slag helps to save energy, decreases graphite electrode consumption and enhance the refectory service life. Foamy slag quality is defined by its chemical and physical properties such as basicity and viscosity. By control of additives injection rate, good quality foamy slag could be obtained. In our research, a neural network model is designed and trained to estimate the rate of additives injection for achieving good quality foamy slag in the DRI based steel making, results of the research is presented in this paper. 
Due to effect of multiple variables in making of fully foamy slag, the “Multi Task Learning Model” has been considered. The neural network has been designed and trained with a specific architecture using real historical data which are related to those heats that had good basicity of slag. The data are belong to a steel making plant in Iran.
There are three key factors that indicate quality of the foamy slag, one is noise made during melting process, another is slag height and the last is the total harmonic distortion (THD)
Therefore, the network is trained by:
-	Additive weights
-	THD of voltage and current
-	the 7th current harmonic.

The foamy slag process has been simulated using this model. The result shows that the trained multi task neural network improved remarkable percent of additive injection rate compared with the operator decision.



CLASS:PUBLIC
DTSTART:20230614T115000
DTEND:20230614T121000
END:VEVENT
END:VCALENDAR