Control of a Separation Column for 18O Isotope Production

In the paper, a method to control the strong nonlinear process associated to a separation column used for the 18O isotope production, is presented. In order to obtain an accurate process model, neural networks are used. The proposed control strategy is analyzed both in the case when the main disturbance signals occur and they do not occur in the system. In order to improve the control system performances, the theory of fractional-order controllers is applied, combined with an adaptive algorithm. For a more efficient disturbances effect rejection, the reference model of the separation process is used.