Impact of model parameters in RANS modeling of urban wind flow: the case study of Livorno city

In the present study the relative importance of several model parameters related to RANS simulation of urban wind flow was investigated, using a district of Livorno city (Italy) as a case study for which earlier wind tunnel tests were performed. CFD simulations were performed at the same reduced scale (1:300) and for the same inflow wind direction α = 240° as the wind tunnel tests. The impact of three model parameters was investigated: the level of geometrical simplification (two levels of detail), inflow conditions (two profiles) and turbulence model (four turbulence models). Four validation metrics were used to quantify the impact of each considered parameter on the calculated mean wind velocities near the ground: Fractional Bias (FB), Normalized Mean Square Error (NMSE), correlation coefficient (R), and the fraction of data within a factor of 1.3 (FAC1.3). The geometrical simplifications were shown to have a much larger effect on the results than inflow conditions and turbulence modeling approach, both in terms of R and FAC1.3.