Predictive Control with Disturbance Forecasting for Greenhouse Diurnal Temperature Control

Abstract

AbstractThis work is devoted to analyze the Generalized Predictive Control (GPC) strategy considering measurable disturbances. This well known control architecture is extended using Double Exponential Smoothing (DES) technique to perform future disturbance estimation. The disturbance models are obtained, validated, and embedded within a GPC controller to compensate for future disturbances. The proposed system is compared with a typical GPC without feedforward action, a GPC with feedforward considering constant disturbances in the future, and a GPC with feedforward taking the original real data in the future. The proposed control scheme was tested by simulation of a greenhouse inside temperature control. The obtained results show that the GPC with disturbance forecasting provide improved behavior that standard techniques.