Reactor furnace control using artificial neural networks and genetic algorithm

Reactor furnace control design is described in this paper. Reactor furnace, as controlled plant, is significantly nonlinear system and control objectives are strict. Thus, special control technique has to be applied. In this case, discrete PID controller tuned by artificial intelligence techniques is chosen. After reactor furnace description and first principle model derivation, the control method is closely defined. Then, nonlinear neural model of the furnace is designed and control loop is connected. Lastly, control performance is shown and compared to performance derived by conventional control method.