Identifying Weakly Connected Subsystems in Building Energy Model for Effective Load Estimation in Presence of Parametric Uncertainty

Abstract

It is necessary to estimate the expected energy usage of a building to determine how to reduce energy usage. The expected energy usage of a building can be reliably simulated using a Building Energy Model (BEM). Many of the numerous input parameters in a BEM are uncertain. To ensure that the building simulation is sufficiently accurate, and to better understand the impact of imprecisions in the input parameters and calculation methods, it is desirable to quantify uncertainty in the BEM throughout the modeling process. Uncertainty quantification (UQ) typically requires a large number of simulations to produce meaningful data, which, due to the vast number of input parameters and the dynamic nature of building simulation, is computationally expensive. Uncertainty Quantification (UQ) in BEM domain is thus intractable due to the size of the problem and parameters involved and hence it needs an advanced methodology for analysis. The current paper outlines a novel Weakly-Connected-Systems (WCSs) identification-based UQ framework developed to propagate the quantifiable uncertainty in the BEM. The overall approach is demonstrated on the physics-based thermal model of an actual building in Central New York.