Advanced control strategies for the modulation of solar radiation in buildings: MPC-enhanced rule-based control

Abstract

The present work is aimed at exploring the potentials of extracting control rules of a smart glazing from an optimal control strategy obtained by means of an ideal model predictive controller (MPC). To this sake an ideal deterministic MPC (Model Predictive Control with ideal prediction of disturbances), minimising total energy use, is devised for the control of a smart glazing for a South-oriented enclosed office space in London. Secondly, a data mining-based method is adopted to extract sets of rules from the simulated MPC-controlled data set. Finally, these rules are compared with the ideal MPC performance, and with reference threshold-based control rules.