BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
BEGIN:VEVENT
ORGANIZER;CN=ESTAD 2023:mailto:info@metec-estad.com
LOCATION:Room 19b
SUMMARY:Exploratory study on the relationships between steel skull and operation data in the RH process using machine learning
DESCRIPTION:In order to improve the efficiency of the decarburization reaction and control the ingredient of molten steel in the Ruhrstahl Heraeus (RH) degassing process, it is necessary to maintain an appropriate vacuum level and keep the refractory in contact with the molten steel clean. However, the temperature difference between the molten steel and refractories causes the steel skull and the solidified slag to adhere to the refractory, affecting the steel component deviation and steel product quality. Due to the absence of a quantitative measurement method for the present amount of steel skull in the RH process, it is difficult to establish an accurate relationship between operation variables affecting the precise element component control of molten steel. In addition, this leads to operation that relies on know-how of individual operator during the RH process. In this study, the quantitative measurement and its evaluation of the steel skull were conducted for improving the temperature control and adjustment of molten steel component during the RH operation. The amount of steel skull was estimated using cross-section images of vacuum vessel before and after the RH operation plant. The correlation between quantitative amount of steel skull and actual RH operation data including variables such as steel component, temperature and time was analyzed statistically technique machine learning. The steel skull amount as a dependent variable was categorized, and the machine learning modelling using the Random Forest (RF) and the Extra Tress (ET) algorithms was performed using approximately 620 data of RH operation. The obtained results have been discussed in context of the knowledge of steel making metallurgy. Physicochemical analysis about the change in the amount of steel skull and correlation between RH operation variables revealed the effectiveness of the proposed machine learning approach.
CLASS:PUBLIC
DTSTART:20230615T113000
DTEND:20230615T115000
END:VEVENT
END:VCALENDAR