Modeling and neuro control for multicomponent nonideal distillation column

The control problem of a multicomponent distillation column is discussed in this paper. A new dynamic mathematical model of nonideal distillation process is derived. Since it contains an implicit recursive procedure, the traditional identification methods cannot be applied, a differential neural network is used to identify this system. Based on the neural identifier, a local optimal neuro controller is proposed. Numerical simulations, concerning a distillation column with 5 components and 15 trays, illustrate the effectiveness of the suggested approach.