Dp 458 en
Modeling and predictive control of buildings
Author: Žáčeková Eva
Due to a huge amount of energy put on the heating in buildings, it is important to make the building heating process more efficient in order to lower the costs. Conventional control methods are not able to take into account fast weather changes, control requirements and others. Many of these drawbacks can be eliminated by making use of predictive controller (MPC).
Besides plenty of advantages like an ability to include given constraints directly into the control law or simple aplicability with MIMO systems, it has also some disadvantages, for examples it needs a mathematical model of system describing its behaviour as accurately as posible in order to work properly. In this thesis, increased attention is paid to searching for mathematical model of system suitable for use with predictive controller. For construnction of Czech Technical University (CTU) in Prague building model, both classical methods and methods minimizing multistep ahead prediction error providing with a model with good prediction properties are used. The chosen model is then employed with desinged predictive controller for temperature control in CTU building rooms.
Eva Žáčeková, tel: +420 776 724 611,mailto:firstname.lastname@example.org