a jahanmiri - department of chemical engineerign university of shiraz
m aghajani - ms.student of chemical

this paper demonstrates that neural networks can be used effectively for the identification of nonlinear dynamical systems two different kinds of neural networks multi layer perceptron MLP and radial basis function RBF networks have been developed to infer compostions of a ternary distillation column from temperature measurements the reliability of the developed setimators is discussed with respect to disturbance on steam reflux rate feed flow rate and feed compositions the estimation results of MLP and RBF networks are represented respectively and some comparisons are carried out

neural network,identification , distillation column