Paper Conference

Proceedings of ASim Conference 2016: 3rd Asia conference of IBPSA-China, Japan, Korea


Building Energy Model Reduction using Principal Component Analysis and Affinity Propagation Clustering of Thermal Zones

Zixiao Shi, William O'brien

Abstract: This paper introduces a building energy model reduction method by using one exemplar zone to represent a group of similar thermal zones. The procedure involves using principal component analysis to model and capture the thermal behaviors of the zones, and then use affinity propagation clustering technique to group similar zones together and identify exemplar zones for the clustered groups. A tool has been developed to allow the process to be performed automatically. A case study discussed in this paper demonstrates the proposed method has produced a reduced energy model that allows a magnitude faster simulation than the original model while still maintains the resulting energy consumption estimation within a reasonable range.
Keywords: Model Reduction, Machine Learning, Building Energy Simulation, BIM