Neural-network-based composite disturbance rejection control for a distillation column

Abstract

Binary distillation columns are essentially multi-variable systems with couplings, non-minimum phase characteristics, model mismatches and various external disturbances. To get the desired top (distillate) and bottom product composition, a composite disturbance rejection control strategy using a radial basis function network (RBFN) is proposed in this paper. The composite controller includes neural network inverse controller (NNIC) and neural network disturbance observer (NNDOB) both using the inverse model of system which is identified by the RBFN. The stability of the identified inverse model is proved, and a rigorous analysis is also given to show why the NNDOB can effectively suppress the disturbances. Performances of the proposed scheme are compared with PID and NNIC without disturbance compensation in three cases by simulation studies. The simulations demonstrate the feasibility, effectiveness and disturbance rejection property of the proposed method in controlling the product composition of the binary distillation columns.