[Back]


Talks and Poster Presentations (with Proceedings-Entry):

M. Killian, M. Kozek, A. Leitner, R. Goldgruber:
"Optimization of smart homes using mixed-integer quadratic-programming and occupancy predictions";
Talk: 8th International Symposium on Energy, Aberdeen, Scotland, UK (invited); 2018-08-06 - 2018-08-09; in: "Proceedings of 8th International Symposium on Energy", North Sea Conference & Journal, (2018), 1 pages.



English abstract:
Buildings in Europe are responsible for approximately two-thirds of the total primary energy consumption. Therefore, concerns over changing climatic conditions (i.e. global warming, depletion of ozone layer, etc.), energy security, and adverse environmental effects are growing among governments, researchers, policy makers, and scientists in developed as well as developing countries. The building energy consumption in Europe amounts to 40 % - 42 %, where 35 % - 40 % are related to CO2 emission. However, the potential in energy savings are estimated to be between 27 % - 30 %. In addition, today´s heating and cooling makes up almost half of the mentioned energy demand in buildings to guarantee user comfort, thus, intelligent building automation solutions are of increasing importance to ensure energy reduction in developed countries.
Model predictive control (MPC) optimally utilizes predictions of future disturbances (ambient temperature, solar radiation, occupancy), and it is the ideal control strategy to deal with conflicting optimization goals, which naturally arise in building automation. Maximization of user comfort and minimization of energy cost are the typical optimization problems in modern smart homes, but also in the smart home energy management system different goals occur such as maximization of usage of alternative energy sources and the resulting green footprint and minimizing monetary cost. Furthermore, MPC is perfectly suited for including thermal and technical constraints in the optimization criterion, considering large time constants (due to thermal inertia and good insulation), and decoupling of multi-variate control problems, thus rendering MPC a powerful control scheme. MPC also enables integration of smart homes into future smart grids.
This research focuses on a smart home including the thermal, the electrical and the grid model. Furthermore, smart appliances are modeled as well as a battery storage. Moreover, a mixed-integer quadratic-programming (MIQP-MPC), an occupancy prediction and a resulting user-adaptive constraint handling, and a flexible MPC tuning tool have been developed and trusted in different simulation studies. Because of the MIQP global control scheme the overall global optimum is guaranteed. The developed concept is flexible, effective, and user-friendly, which is a necessary condition in smart homes.

Keywords:
smart home, energy management system, energy efficiency, model predictive control, smart appliances, mixed-integer quadratic-programming

Created from the Publication Database of the Vienna University of Technology.