Title: Hybrid Models for Steelmaking and Casting Applications
Presenter: Ruibin Wang (PhD candidate, 2nd seminar)
Supervisor: Prof. K. Chattopadhyay
Abstract:
Steel production involves a series of processes that convert iron and steel scrap into end products for industrial application. Over the past decade, steelmaking industry has faced continuous economic and environmental challenges. As a result, for fulfilling the steel product quality requirement and ensuring maximized process efficiency, strict monitoring for the Basic Oxygen Furnace (BOF) and continuous casting processes is needed.
Exist control models for the BOF process are mostly developed based on thermodynamic principles or by deploying advanced sensors. In the present study, a novel hybrid method for endpoint temperature, carbon, and phosphorus based on heat and mass balance coupled with data-driven technique is proposed. Three types of static models were established, firstly, theoretical framework based on user specified inputs were formulated based on mass and energy balance; secondly, artificial neural networks (ANN) were developed for end-points predictions; finally, the proposed hybrid model was established based upon exchanging outputs among theoretical models and ANNs. The application of the hybrid model can successfully improve the prediction accuracy of theoretical framework and generalization of machine learning models.
During the continuous casting process, the clogging of submerged entry nozzle (SEN) is a critical issue that adversely affects final product quality and process productivity. In order to impose effective monitoring and control over the continuous casting process, a quantitative index was formulated to quantify the magnitude of SEN clogging and erosion for ultra-low carbon, low carbon, medium carbon, and calcium treated grades. Three critical index values are defined to represent the clogging event, erosion event, and critical casting condition. Long short-term memory network was established based on the quantitative index in the past four minutes to predict that in the future 48 seconds. The models can identify most of the critical casting conditions and erosion incidents for all steel grades. In the production setting, operators can take corresponding actions when critical conditions are predicted in order to prevent the possible occurrence of clogging.