Process-time Optimization of Vacuum Degassing Using a Genetic Alloy Design Approach.

David Dilner,Malin Selleby,Huahai Mao,Sybrand Van Der Zwaag,Qi Lu,Wei Xu

doi:10.3390/ma7127997

Abstract

This paper demonstrates the use of a new model consisting of a genetic algorithm in combination with thermodynamic calculations and analytical process models to minimize the processing time during a vacuum degassing treatment of liquid steel. The model sets multiple simultaneous targets for final S, N, O, Si and Al levels and uses the total slag mass, the slag composition, the steel composition and the start temperature as optimization variables. The predicted optimal conditions agree well with industrial practice. For those conditions leading to the shortest process time the target compositions for S, N and O are reached almost simultaneously.

Highlights

Several approaches to predict materials properties have been developed over the past decades.Among those the CALPHAD-method [1,2], an approach for compuational thermodynamics, is one of the most widely applied
The difference in slag amount, is expected since it is known that a higher slag content promotes desulfurization, see, e.g., a computational fluid dynamics (CFD) study of vacuum degassing [17]
The fact that industrial practice suggests higher pressure indicates that the model used cannot fully reproduce the true nature of vacuum degassing

Summary

Introduction

Several approaches to predict materials properties have been developed over the past decades.Among those the CALPHAD-method [1,2], an approach for compuational thermodynamics, is one of the most widely applied. A similar approach utilizing mesh adaptive direct search (MADS) incorporated in the FactSage system [9] has been developed by Gheribi and Pelton et al, which may be applied to calculate liquid surfaces [10,11] as well as optimizing the composition of magnesium alloys [12] and the blast furnace process [13]. These approaches in the field of materials science preceded the newly started initiative by the U.S

Objectives

Results

Conclusion