In primary metallurgy, the converter process is an important stage in the production of steel from iron ore. The aim of the process is to remove carbon and several minor elements from iron melt by blowing oxygen on top of the melt. The scrap is used mainly to lower the furnace temperature. In a typical converter the iron bath is mixed by inert gas injection through nozzles placed at the bottom of the furnace.Converter process models aim to predict the temperature and compositional changes in the metal and slag phases during the oxygen blow. Process models are usually very simple and compute the results in real time. Real time capability is important for online process control use. The purpose of this thesis is was to improve the numerical modelling of the phenomena that takes place inside the steel converter, with an eye towards real-time applicability. The thesis developed new algorithms and numerical methods for computing gas/liquid two-phase flow, chemical reactions between multiple phases, as well as scrap melting. The models for these three phenomena were divided into three sub-models, which were tested independently. A theoretical basis for combining the sub-models was developed too. All of the models were developed to work in real-time. The approximations used are readily applicable to many other high temperature processes. The scrap melting model was developed to work in 1-D. The main idea was to predict the crust formation when liquid iron solidifies on the cold scrap, and then the melting of the solid material. The phase changes in the solid material are taken into account with a moving numerical grid. The scrap melting model was validated against melting measurements done in a crucible for three different melt temperatures. The numerical results compared well with the measurements.The 2-D two-phase computational fluid dynamics model was developed to predict the flow field and mixing of the iron melt due to the bottom gas stirring. The algorithm was developed to conserve liquid mass with computational economy, which was a feature missing in the literature. The model was validated with three different water models, one of which was similar to an industrial-scale steel converter. The chemical reaction model was based on computing chemical thermodynamic equilibrium via the Gibbs energy minimisation procedure for a reaction volume, which exists in the interfacial region of the main phases. The evolution of the chemical reactions are then determined by the mass transfer between the reaction volume and the bulk phases. Computing the equilibrium allows the modelling of simultaneous reactions and predict possible saturation of the components. The model was validated against measurements done in an industrial-scale steel converter. Accuracy was improved over existing literature models.
|Translated title of the contribution||Multifysikaalinen työkalu teräskonvertteriprosessiin|
|Publication status||Published - 2018|
|MoE publication type||G5 Doctoral dissertation (article)|
- steel converter
- numerical modeling