Eingang zum Volltext in OPUS
Hinweis zum Urheberrecht
Dissertation zugänglich unter
RANS versus LES models for investigations of the urban climate
RANS versus LES Modelle für Stadtklimauntersuchungen
Fock, Bjoern Hendrik
Dokument 1.pdf (26.018 KB)
Freie Schlagwörter (Englisch):
LES , turbulence , convection , urban climate
38.80 , 38.81 , 38.82 , 38.87 , 38.84
Schlünzen, K. Heinke (Prof. Dr.)
Tag der mündlichen Prüfung:
Kurzfassung auf Englisch:
This work contributes to the capabilities of turbulence permitting simulations for investigations of the urban climate, by extending and testing the mesoscale model METRAS for large eddy simulations. Model experiments focus on convective boundary layers above homogeneous and heterogeneous surfaces. The heterogeneous surfaces represent an urban area. The performed simulations are used to test METRAS-LES and to investigate topographic impacts on convective boundary layers above the inner-city and harbor area of Hamburg.
A new time integration scheme for the dissipation term in the TKE-equation is presented. A method to control simulations by prescribed heat fluxes is newly implemented into the model. Model intercomparisons between METRAS-LES and the three models METRAS-RANS, METRAS-PCL and PALM contribute to the model testing and demonstrate the difference between RANS and LES solutions. Analyzed sensitivity studies cover the influence of vertical grid stretching, numerical methods and details of the subgrid scale model. Metrics used to discuss model results include spectra, percentile differences and characteristic heights derived from domain averaged variance and covariance profiles. New surface class mappings have been defined for ATKIS and CORINE data, which are used for the simulations above heterogeneous areas. An additional analysis of annual and diurnal cycles of measured water temperatures is provided.
The model experiments show, that free convective boundary layers can be reasonable simulated with METRAS-LES. An effective model resolution of 6 – 8 Dx could be found for the model results. Grid stretching is acceptable, as long as the area of interest remains well resolved. The stability dependent formulation of the length scale in the used SGS model seems to be of minor importance, compared to other influencing parameters. The TKE parameterization in the surface layer, which is only suitable for flux matching in METRAS-RANS, can be used in METRAS-LES to tune the diffusivity of the model. Simulations with METRAS-LES do not only work for prescribed heat fluxes, but also with the surface energy balance. This allows to study diurnal cycles of convective boundary layers above heterogenic urban surfaces. Simulations of convective boundary layers above Hamburg show a clear response to the water bodies of Elbe and Alster. Experiments with modified topography are used to investigate the influence of orography and surface heterogeneity. The sensitivities on modified topography are relative small for most experiments. The largest impacts, which are even visible by reduced domain averages of 10 m wind speeds and temperatures, are found for redistributing the water areas of Elbe and Alster, randomly within the complete model domain.