GEOMETRIC ANALYSIS OF THE GLOBAL STABILITY OF LINEAR INVERSE-BASED CONTROLLERS FOR BIVARIATE NONLINEAR PROCESSES

Many nonlinear processes have highly structured model mismatches that make a controller, based on an inverse of a linear nominal model, more robust than is predicted from the theories assuming unstructured, norm-bounded model mismatch. This work focuses on the robust stability problem of inverse-based controllers for 2 x 2 processes where process nonlinearities are the dominant source of model mismatch. Some geometric criteria are proposed to examine whether these linear controllers can stabilize a nonlinear process over a wide range of operations. High-purity distillation column control is used as an example of physical processes with severe nonlinearities and highly structured model mismatch.