Abstracto

Controlling and modeling phases of distillation column using artificial neural networks

Mohammad M.Zarei, Jafar Sadeghi, Fariba Zarei

In this paper is described the choice of the control system design for a binary distillation column. The column has been formed dynamically. Using artificial neural networks (ANN) in system control design, the effects of disturbance on the column has been rejected to modeling ANN. These networks are used to model complex and non-linear processes and have the potential to solve some types of complex problems, where traditional methods wonÂ’t answer properly. Using dynamic simulation, the proper educational, testing and validation data for designing neural network is yielded. Modeling the system is done by multi layer perceptrons Levenberg-Marquardt algorithm. Finally, according to the error result values, acceptable errors for neural network are presented.