Paper Conference

Proceedings of eSim 2014: 8th Conference of IBPSA-Canada


Application of data mining techniques for energy modeling of HVAC sub-systems

Mathieu Le Cam, Radu Zmeureanu, Ahmed Daoud

Abstract: The Building Automation Systems (BAS) installed in commercial and institutional buildings collects a very large amount of data. These measurements present a gold mine of information which could be used for better understanding of the actual building operation and performance, or for fault detection. This study presents practical use of data mining to extract information from BAS measurements, for development of inverse models of building energy performance. The case study results show that the target variable (the building demand for chilled water) depends on the supply air temperature, humidity and enthalpy in the air-handling unit (AHU), the cooling coil valve modulation in both AHUs, and outside air enthalpy. About 75% of the variability of the dataset can be explained by using only four principal components calculated from the original dataset. The clustering analysis revealed that four daily profiles could describe the chilled water daily use in summer and autumn seasons.
Pages: 96 - 109