Building system performance diagnosis and optimization based on data mining techniques

Buildings are becoming not only energy-intensive, but also information-intensive. Today's building automation system (BAS) has provided an enormous amount of data about the actual building operation. Valuable insights could be gained from such data. However, due to the data complexity and the lack of advanced analytic tools, only limited and rather simple applications have been found. Data mining (DM) is a promising technology which has great efficiency and effectiveness in discovering hidden knowledge from massive data sets. This study investigates the utilization of DM in analyzing massive BAS data for enhancing building energy efficiency. The DM-related research in the building field is firstly reviewed and then the challenges of practical applications are discussed. A generic DM-based analysis framework is proposed. The framework is applied to analyze the building operational data retrieved from the tallest building in Hong Kong. Two case studies are presented to show the capability of DM in developing robust energy prediction models, identifying building operating behaviors, and evaluating operational performance. The results show that, DM, together with domain knowledge, could be very powerful in the knowledge discovery in massive building operational data and valuable for enhancing the building energy efficiency.