Robust Control of Variable Flow Processes

Abstract

Models of continuous flow processes with variable parameters are studied and methods developed for their robust control. Such processes have proven invariant under variable flow and volume, if they are presented as functions of composite variables which have been introduced recently. Their finite impulse response (FIR) models are obtained e.g. by physical testing or identification, and expressed in terms of discrete, equidistant values of the appropriate variable. In the present study, such a model of a time-variable process with a long time delay is considered a high-order model; this is subsequently reduced to a low-order state-space model by means of the optimal Hankel-norm approximation. The robust feedback control is developed in terms of the same variable on the basis of the low-order model. The application of the control algorithm to the simulated process proves the approach superior to such conventional and robust control methods which have the regular time as their independent variable under unsteady conditions.