Using data mining in optimisation of building energy consumption and thermal comfort management

Abstract

Performance monitoring using wireless sensors is now common practice in building operation and maintenance and generates a large amount of building specific data. However, it is difficult for occupants, owners and operators to explore such data and understand underlying patterns. This is especially true in buildings which involve complex interactions, such as ventilation, solar gains, internal gains and thermal mass. Performance monitoring requires collecting data concerning energy consumption and ambient environmental conditions to model and optimise buildings' energy consumption. This paper details the use of data mining techniques in understanding building energy performance of geothermal, solar and gas burning energy systems. The paper is part of an outgoing research into optimisation of building performance under hybrid energy regimes. The objective of the research presented in this paper is to predict comfort levels based on the Heating, Ventilating, and Air Conditioning (HVAC) system performance and external environmental conditions. A C4.5 classification methodology is used to analyse a combination of internal and external ambient conditions. The mining algorithms are used to determine comfort constraints and the influence of external conditions on a building's internal user comfort. To test the performance of classification and its use in prediction, different offices, one to the south and the other to the north of the building are used. Classification rules being developed are analysed for their application to modify control algorithms and to apply results to generalise hybrid system performance. The results of this study can be generalised for an entire building, or a set of buildings, under a single energy network subject to the same constraints.