Author(s):
1. Vesna Ranković, Faculty of Engineering, University of Kragujevac, Sestre Janjić 6, 34000 Kragujevac, Srbija 2IMW Ins, Serbia
2. Milorad Bojic, Faculty of Engineering, University of Kragujevac, Sestre Janjić 6, 34000 Kragujevac, Srbija 2IMW Ins, Serbia
3. Jasna Radulovic, Fakultet inženjerskih nauka Univerziteta u Kragujevcu, Serbia
4. Danijela Nikolic, Faculty of Engineering, University of Kragujevac, Sestre Janjić 6, 34000 Kragujevac, Srbija 2IMW Ins, Serbia
5. Jasmina Skerlic, Fakultet inženjerskih nauka Univerziteta u Kragujevcu, Serbia

Abstract:
Model-based predictive control is a very modern powerful control strategy that uses a model of the plant to predict its future behaviour and has been in focus of researchers in the area of buildings energy management. The aim of this paper is to develop and analyze NARX(Nonlinear AutoRegressive with eXogenous inputs) structure for the modeling of the building thermal behavior. A database was generated using simulation in EnergyPlus software. The NARX identification model was designed using the MATLAB System Identification Toolbox. The input variables analyzed in this paper are: outdoor temperature, heating or cooling power, and direct solar radiation. Simulation results demonstrate that the proposed nonlinear structure can be effective identification tool for development of nonlinear buildings predictive control.

Key words:
building, NARX, identification, thermal behaviour

Thematic field:
Energy and Thermal Engineering

Date of abstract submission:
04.03.2015.

Conference:
12th International conference on accomplishments in Electrical and Mechanical Engineering and Information Technology (DEMI 2015)

Final paper