Filtered Smith Predictor with nonlinear model applied to a solar field

Abstract

Model predictive control such as Generalized Predictive Control (GPC) combined with feedback linearization (GPC-FL) has shown to be a successful methodology to regulate the outlet temperature in a solar field. In addition, the benefits of input-output feedback linearization have been improved by a Dead Time Compensator (DTC) to avoid unstable or highly oscillatory responses caused by plant-model mismatch. In this paper, a Filtered Smith Predictor (FSP) controller is proposed using a nonlinear model (FSP-NLM) of the plant instead of a nominal linear one. Simulation studies compare DTC-GPC-FL with FSP-NLM and advantages of each one are explained.