Comparison of Sensitivity Analysis Methods in Building Energy Assessment

TLDR

Sensitivity analysis is an important tool in building energy assessment to determine the key factors influencing energy use or carbon emissions for buildings. The study compares the characteristics of four global sensitivity analysis methods—SRC, Morris, extended FAST, and TGP—in building energy assessment. A retail building in Harbin, China, served as a case study to illustrate the advantages and drawbacks of these four methods. The TGP method outperformed the others in accuracy and computational cost, though it requires more time to compute sensitivity indices; the authors recommend using at least two fundamentally different methods for robust results.

Abstract

Sensitivity analysis is an important tool in building energy assessment to determine the key factors influencing energy use or carbon emissions for buildings. This research is focused on comparing the characteristics of four global sensitivity analysis: SRC (standardized regression coefficient), Morris design, extended FAST (Fourier Amplitude Sensitivity Test) and TGP (treed Gaussian process) method. A retail building located at Harbin (China) is used as a case study to demonstrate the advantages and drawbacks for these four methods. The results indicate that the TGP method (one of meta-modelling approaches) is the best choice in terms of both accuracy and computationally cost. Note that the TGP method needs more time to calculate the sensitivity index although it needs only moderate time for running building energy models. At least two fundamentally different methods for sensitivity analysis are recommended to be performed to provide more robust results in building energy assessment.