Neural control strategies of a binary distillation column

The ability of neural networks to model arbitrary nonlinear functions and their inverses is exploited for the adaptive control of nonlinear systems. Neural networks which model the plant and its inverse are directly incorporated within the internal model control structure. In addition, a test was made with the open loop control using only the neural model of the plant inverse. Finally, combined structures of conventional controllers (P, PD) with this inverse model were implemented in order to improve the performance of the controlled system. The potential of the proposed methods is demonstrated using the control of the top of a continuous Benzene-Toluene distillation column as an example. The dynamic behavior of that system is obtained by using a complex software simulation