Comparative Analysis of Uncertainty Characterization Methods in Urban Building Energy Models in Hot-Arid Regions

The development of reliable building energy models at the urban scale is crucial for analyzing and optimizing the energy efficiency of cities. The bottom-up physics-based approach has been widely employed in Urban Building Energy Models (UBEMs). However, the uncertainty of input parameters can impact the reliability of UBEM simulation outputs, and very limited studies considered the uncertainty when developing archetype models for UBEMs. While UBEMs typically rely on a traditional deterministic approach, incorporating probabilistic methods can significantly enhance simulation accuracy by accounting for uncertain variables. Probabilistic methods involve characterizing key uncertainties in input data using Probability Distribution Functions (PDFs). Yet, the effect of using different PDF types on UBEM results is not adequately understood, and the literature often assumes uniform distribution. In this study, UBEM is characterized based on three methods. The deterministic approach is used to serve as a baseline, and two different PDF types are used to examine how PDFs impact simulation results when uncertain parameters are present in UBEMs. Latin Hypercube Sampling (LHS) is employed to propagate uncertainty in input parameters in UBEM. The study is conducted on a case study area of the Marina district of Lusail City, Qatar, characterized by a hot and arid climate.