Optimisation of Ferrochrome Addition Using Multi-Objective
Evolutionary and Genetic Algorithms for Stainless Steel Making
via AOD Converter

Abstract

In this work, optimisation study has been carried out between ferrochrome utilisation and end blow chromium content of steel during stainless steel making in AOD converter. The objective of optimisation is to lower the use of ferrochrome and increase the chromium content of stainless steel. A thermodynamic mathematical model is constructed to simulate the whole production process for 204Cu grade stainless steel and by varying the values of the process variables between certain ranges ,100 dataset is generated which are verified using ind ustrial furnace data. Sixteen input process variables and two objectives are accounted during the optimisation work. The process variables used during the optimisation work are oxygen ,argon ,nitrogen blowing rate, initial bath temperature, duration of blowing, chromium and carbon content, weight of ferrochrome .Evolutionary neural network is used to construct two meta models which mapps the objective function to the process variables. These meta models are optimised using predator prey algorithm which leads to formation of pareto frontier consisting a set of nondominant solutions. The paretofrontier addresses the issue of optimum ferrochrome utilisation for stainless steel of specific chromium content. Single variables response analysis is carried out to analyse the correlation among the process variables and objectives.