Uncertainty analysis in building energy simulation: A practical approach

Abstract

Uncertainty analysis in building energy simulation is often carried out with Monte Carlo Analysis. However, there is currently no standard framework for uncertainty modeling in building energy models. In particular, uncertainty quantification is often based on literature and expert judgment with limited attention on the use of measured data. The purpose of this paper is to provide a structured framework for modeling uncertainties in building energy models. A method for selecting probability distributions based on measured data is presented, allowing users to assign probability distributions that better represents actual data distributions. If measured data is not available, the subjective assignment of an appropriate probability distribution would be based on drawings, specifications, literature and past case studies on uncertainty quantification. To facilitate its implementation to building design and retrofit analysis, an application that couples the proposed framework with EnergyPlus is introduced. Demonstrated through a case study, the proposed framework provides acceptable simulation results in comparison with measured gas and electricity consumption.