Clustering-based performance assessment of thermal energy management in buildings

In recent years a growing attention is being paid towards energy efficiency in civil and industrial applications. In order to achieve better efficiency and enhance the energy management strategies, better knowledge is needed on the characterization of energy consumers. Categorization of electrical consumers through clustering procedures has been widely addressed in the literature. However, the categorization of the thermal users has been considered to a lower extent, also because of the difficulty of gathering a sufficient and reliable amount of data. This paper analyzes the behavior of several final users on the thermal side. At first, the energy behavior of the buildings from a thermal point is defined by its energy signature, represented as a straight line derived from measurements gathered in the heating period. Then, the parameters of the energy signature are used within a clustering procedure to create consistent groups of buildings. The results enable the operators to identify salient features of the groups of buildings under analysis, and to monitor the changes carried out in the buildings, leading to possible changes of the group during time.