Large-eddy simulations to define building-specific similarity relationships for natural ventilation flow rates

Natural ventilation can play an important role towards preventing the spread of airborne diseases in indoor environments. However, quantifying natural ventilation flow rates is a challenging task due to significant variability in the boundary conditions that drive the flow. In the current study, we propose and validate an efficient strategy for using computational fluid dynamics (CFD) to assess natural ventilation flow rates under variable conditions, considering the test case of a single-room home in a dense urban slum. The method characterizes the dimensionless ventilation rate as a function of the dimensionless ventilation Richardson number and the wind direction. First, the high-fidelity large-eddy simulation predictions are validated against full-scale ventilation rate measurements. Next, simulations with identical Richardson numbers, but varying dimensional wind speeds and temperatures, are compared to verify the proposed similarity relationship. Last, the functional form of the similarity relationship is determined based on 32 LES. Validation of the surrogate model against full-scale measurements demonstrates that the proposed strategy can efficiently inform accurate building-specific similarity relationships for natural ventilation flow rates in complex urban environments.