Uncertainty and sensitivity analysis of natural ventilation in high-rise apartment buildings

Quantification of natural ventilation rates is an important issue in HVAC system design. Natural ventilation in buildings depends on many parameters whose uncertainty varies significantly, and hence the results from a standard deterministic simulation approach could be unreliable. This study performed uncertainty analysis to predict natural airflow rates. The paper presents relevant uncertainty in model parameters such as meteorological data, building  properties (leakage areas of windows, doors, etc.), etc. Uncertainties of the aforementioned parameters were quantified based on data available from literature and on-site visits. The Monte-Carlo method with Latin Hypercube Sampling (LHS) was used for uncertainty propagation. The CONTAMW was chosen to simulate natural ventilation phenomena in a high-rise apartment building that is typical of residential buildings in Korea. It is shown that the uncertainty propagated through this process is not negligible and may significantly influence the prediction of the airflow rates. In the paper,