Building resolving large eddy simulations and comparison with wind tunnel experiments (Q2462482)

The authors perform large eddy simulation (LES) of a flow past a scale model of a complex building, using two different methods to represent the edifice, both methods being implemented in the same hydrodynamical code (EULAG). The first part represents an introduction concerning the numerical modeling of natural urban flows, and two distinct methods used to represent an edifice: the first one using the \textit{T. Gal-Chen} and \textit{R. C. Somerville} terrain-following coordinate transformation [J. Comput. Phys. 17, 209--228 (1975; Zbl 0297.76020)], and the second one employing the immersd boundary approach. In the second part, the theoretical formulation of the fluid dynamics model is outlined. The nonhydrostatic model EULAG used in this study has been thoroughly documented in the literature, and it admits several optional formulations of equations of motion. Taking into account the small-scale boundary-layer flows, here the authors adopt the classical incompressible Boussinesq approximation. The third part concerns the numerical approximations of governing equations. The calculations use exclusively the second-order-accurate monotone FCT, \textit{P. K. Smolarkiewicz} and \textit{W. W. Grabowski} [J. Comput. Phys. 86, No. 2, 355--375 (1990; Zbl 0698.76100)], flux-form scheme MPDATA. Design of the numerical experiments and the cortesponding results are discussed in the fourth part. The fifth part contains the results and remarks on the numerical simulations which are divided into two groups: one employing the Gal-Chen and Somerville coordinate transformation, and other based on a Cartesian mesh using the immersed-boundary approach. The two solution methods are compared to wind tunnel data collected for neutral stratification. Profiles of the first- and second-order moments at various locations around the model building show good agreement with wind tunnel data. Although both methods appear to be viable tools for LES of urban flows, the immersed-boundary approach is computationally more efficient. The sixth part presents a stably stratified case to compare and contrast to the neutral flow results given in the previous part, pointing out the substantial differences. The concluding remarks and the future perspectives are presented within the last part of the paper.