Using adaptive neuro fuzzy inference system in developing an electrical arc furnace simulator
Farrokh Janabi‐Sharifi, G. Jorjani, Iraj Hassanzadeh · 2006
This paper presents the use of adaptive neurofuzzy inference systems (ANFIS) in simulating the regulator control loop of the electrical arc furnace (EAF). The regulator loop is the core part of steel making EAF, which controls positioning of the electrodes. The non-linearity and complexity of EAF makes it very difficult to use the classical mathematical modeling techniques in building the process simulator. This research shows that, the EAF regulator loop could be modeled with the use of ANFIS as non-parametric modeling method. The effort is extended to put together the different parts of the model in a cascade and come up with a complete regulator loop simulator. The simulator outputs are illustrated beside the actual recorded plant data. The actual data used were acquired and recorded from the EAF of the Gerdau Ameristeel Whitby (GAW) in Ontario, Canada