Energy consumption estimation for building energy management systems (BEMS) is one of the key factors in the success of energy saving measures in modern building operation, either residential buildings or commercial buildings. It provides a foundation for building owners to optimize not only the energy usage but also the operation to respond to the demand signals from smart grid. However, modeling energy consumption in traditional physical modeling techniques remains a challenge. To address this issue, we present a data mining-based methodology, as an alternative, for developing data-driven models to estimate energy consumption for BEMSs. Following the methodology, we developed data-driven models for estimating energy consumption for a chiller and a supply fan in an air handling unit (AHU) by using historic building operation data and weather forecast information. The models were evaluated with unseen data. The experimental results demonstrated that the data-driven models can estimate energy consumption for BEMS with promising accuracy.

Air handling unit (AHU), Building energy management systems (BEMS), Data-driven modeling, Demand response systems, Machine learning algorithms
dx.doi.org/10.1007/978-3-319-62575-1_72
Green Energy and Technology
Department of Mechanical and Aerospace Engineering

Yang, C. (Chunsheng), Cheng, Q. (Qiangqiang), Lai, P. (Pinhua), Liu, J, & Guo, H. (Hongyu). (2018). Data-driven modeling for energy consumption estimation. In Green Energy and Technology. doi:10.1007/978-3-319-62575-1_72