CONTROL AND MODELING STRATEGIES FOR NONLINEAR SYSTEMS WITH TIME DELAYS

New control and modeling strategies are developed for a general class of nonlinear systems with time delays. The proposed models are easy to obtain from a limited amount of experimental data and are general enough to represent a large group of chemical processes. The new sub-optimal control scheme is based on a nonlinear model and a general quadratic performance index. It has a simple feedback structure like most standard controllers but has parameters which vary with changing process conditions.;A new prediction scheme with continuous load estimation is also proposed. The well known Smith Predictor is equivalent to the steady-state version of this new predictor. Results from a simulated first-order nonlinear system show that the new predictor is much better than the Smith Predictor.;Simulation studies for two distillation column models and a first-order example with a nonlinear gain show that the proposed control scheme out-performs a well-tuned conventional PI controller for both setpoint and load changes. The advantage of the new control design techniques is further confirmed by real-time computer control of a pilot-scale distillation column unit in the department.