Identification of characteristic user behavior with a simple user interface in the context of space heating
Adolph, Michael; Müller, Dirk (Thesis advisor); Frank, Martin (Thesis advisor)
Aachen / E.ON Energy Research Center, RWTH Aachen University (2017, 2018) [Book, Dissertation / PhD Thesis]
Page(s): 1 Online-Ressource (xvi, 188 Seiten) : Illustrationen
Increasing the energy efficiency of buildings is a promising approach to reduce a society's energy consumption. This work shows an approach to automatically detect user preferences on room temperature and the thermal behavior of a building solely on feedbacks provided by the user. This information is used to create a temperature schedule for each individual room. If the user is not present, the room's temperature is automatically reduced to save energy. Prior to the user's expected return the temperature is raised again. This automated approach allows the user to save energy without compromising his thermal comfort and the need for complex schedule programming. This work evaluates the suggested base algorithm with simulations conducted in Modelica and a field test. The user characteristics are estimated with three different versions of a learning algorithm. To determine the user's preferences, two different user characteristics are used. The effects of three different building standards are compared. Also different approaches to add a necessary pre-heating period are tested. Additionally, the effects of different heating systems are evaluated. The results are evaluated by thermal comfort, energy consumption and the number of feedbacks, i. e. the user's effort. The field test puts the algorithm in a real life scenario and tests its general functionality and its ease of use, using sensor data and a questionnaire. It is found that energy savings of 10-20 % are possible without sacrificing the user's thermal comfort. All this is possible with an easy to use system that relies only on occasional qualitative feedback.