Occupant behavior and air conditioning usage revealed from sensor fusion applying the k-means clustering method

Abstract

The knowledge of the occupant’s behavior in a building allows the evaluation of the occupant’s comfort in it since it takes into consideration aspects of his surroundings or environment that affect them directly, as well as the consideration of entrance/exit in the room and the energy consumption, which allows the evaluation of improvement alternatives in terms of building design. In this study, the k-means algorithm was implemented on data collected (temperature, relative humidity, carbon dioxide) in a room of a two-story residence for one year. The results show that carbon dioxide data is the best for detecting occupant presence, however, all three types of variables were able to detect the use of air conditioning in the case study.