Dynamic process modeling and simulation of electric arc furnaces

The adverse conditions at the electric arc furnace make experimental investigations and parameter variations during the melting process difficult and to some extend are also associated with high costs. For these reasons a dynamic process model is used at the IOB for theoretical investigations and for the optimization of the electric arc furnace process. In contrast to CFD and FEM simulation methods and due to simplifications and a holistic view on the process the dynamic process model delivers results faster with less time required for computing. It also offers a high degree of application versatility:

  • Analysis of different process operation modes
  • Optimization of process control parameters
  • Investigation of alternative EAF equipement and EAF designs
  • Soft sensing
  • Operator training

The analytical process model is implemented using MATLAB and is based on fundamental thermodynamic and physical equations. The model published by Logar, Dovžan und Škrjanc (2012) forms the basis and has been further developed, especially with regard to the gas phase.

The process model consists of separate modules in which the energy distribution and heat transfer as well as relevant chemical reactions and mass transfers are calculated. By means of first order differential equations mass and temperature changes of the single phases (scrap, melt, slag, …) are calculated observing the laws of conservation of energy and mass.

On average the process simulation is carried out within one to three minutes depending on the complexity of the mode of operation. Due to the implementation of the model for parallel computing, depending on the number of physical CPU cores available, multiple melts can be simulated at once.

The time dependent results of the simulation provide the temporal development of the temperature of all phases and zones of the model, masses and chemical compositions as well as heat flows during the melting process.

The process model has been validated using process data of two industrial electric arc furnaces. In its next stage of development the model can be used to investigate alternative modes of operation. The simulation will be adapted in way, that within given boundaries of mass and energy inputs an automatic control of the process by the simulation itself will be possible. Due to parallel computing it will be possible to simulate and evaluate different modes of operation and to optimize the melting process within a short period.