Computational Fluid Dynamics for urban physics: Importance, scales, possibilities, limitations and ten tips and tricks towards accurate and reliable simulations

CFD simulation of wind flow over natural complex terrain: Case study with validation by field measurements for Ria de Ferrol, Galicia, Spain

“…This is demonstrated by both general review papers (e.g. Murakami 1993, Murakami et al 1999, Stathopoulos 1997, 2002, Baker 2007, Blocken et al 2011, Moonen et al 2012, Meroney and Derickson 2014, Blocken 2014, 2015 and review papers on topics of particular interest, such as dispersion of pollutants (e.g. Meroney 2004, Canepa 2004, Tominaga and Stathopoulos 2007, Di Sabatino et al 2013, natural ventilation of buildings (e.g.…”

“…A distinction is made between two types of roughness (Blocken et al 2007a, Blocken 2015: (1) the roughness of the terrain that is included in the computational domain and (2) the roughness of the terrain that is not included in the computational domain. The knowledge of the roughness of the terrain that is situated outside the computational domain is important because it determines the shape of the inlet profiles of mean wind speed and turbulence properties.…”

“…Therefore, CFD verification and validation are imperative, and for this purpose, extensive sets of best practice guidelines for CFD in wind engineering have been developed in the past 15 years (e.g. Casey and Wintergerste 2000, Franke et al 2004, 2007, Britter and Schatzmann 2007, Tominaga et al 2008, Blocken and Gualtieri 2012, Blocken 2015. It should be noted that validation requires high-quality experimental data, either wind-tunnel data or field data, and these data in turn need to satisfy certain quality criteria (Schatzmann et al 1997, Schatzmann andLeitl 2011).…”

Low-Reynolds number mixing ventilation flows: Impact of physical and numerical diffusion on flow and dispersion

“…In addition, several previously published guidelines stated that a grid-sensitivity analysis should be conducted to make sure the results obtained are (nearly) grid independent and do not suffer from excessive amounts of numerical diffusion (e.g. Jones and Whittle 1992;Chen and Srebric 2002;Chen and Zhai 2004;Roache 1997;Li and Nielsen 2011;Blocken 2015). Sørensen and Nielsen (2003) stated that-at that time-obtaining a grid-independent solution for 3D cases was very difficult due to the lack of computational resources.…”

“…As discussed earlier in several publications addressing oscillatory convergence (e.g. Blocken 2015), an "average" solution is obtained by averaging the results over a sufficient amount of iterations (difference when averaging longer becomes less than 1%), over an interval of 1,000-10,000 iterations, depending on grid size, discretization scheme and inlet TI.…”