Finding Good Trade-offs Between Energy Consumption and Occupant's Comfort in Smart Buildings

In buildings, classical heating controllers are based on user-preferred settings of the indoor temperature at particular times of day. The controllers alter the heating or cooling to achieve a desired temperature. Since comfort also relates to other factors, such as human activity rate and indoor air humidity, we developed a novel comfort regulator that computes the value of indoor set-point temperature to achieve the desired comfort dynamically. In this way, finding the best control strategies corresponds to identifying the most user-preferred balances between energy consumption and comfort experience. The second novelty is introduced through a new occupancy-prediction algorithm that improves the comfort experience when a home becomes occupied, and the expert rules that efficiently reduce unnecessary energy consumption during periods when the home is not occupied. Experimental results for a specific building are presented in a comfort/energy consumption space to demonstrate how the Pareto frontier is influenced by (i) different versions of control strategies and (ii) different comfort threshold values. The experiments show that the proposed control system achieved a better comfort experience with small increases in energy consumption compared to a reference approach - that is, a reactive control system that triggers based on occupancy events only.