Neural Network Control for a Batch Distillation Column


The present work deals with studying the dynamic behavior of a batch distillation column and implemented two types of control strategies for the separation different types of binary systems. The model was derived and then simulated using "MATLAB" program. The experimental data of dynamic behavior were to tune the parameters of PID controller and developed the training of neural networks controller by using supervised learning algorithms. The simulation results show a qualitatively acceptable behavior. This study shows also that the response of PID controller was oscillatory behavior with high offset value while neural network controller gave less offset value and less time to reach the steady state. In general, a good improvement is achieved when the neural network controller is used compared with PID control.