Effects of increased geometric complexity on the comparison between computational and experimental simulations

• Published in: Journal of wind engineering and industrial aerodynamics 1998-02, Vol.73 (2), p.159-179
• Main Authors: Byrne, C.E.I., Holdo, A.E.
• Format: Article
• Language: English
• ISSN: 0167-6105; 1872-8197
• DOI: 10.1016/S0167-6105(97)00284-5

CFD studies and methods are generally validated against idealized flow configurations [V.C. Patel, W. Rodi, G. Scheuerer, Amer. Inst. Aeronaut. Astronaut. 23 (9) (1984); W. Rodi, The prediction of free turbulent boundary layers by use of a two-equation model of turbulence, Ph.D. Thesis, University of London, 1972]. Whilst it may be argued that most industrial flows consist of a combination of idealized flow types, the interaction between these flow types is not always well understood. Industrial flows are affected by varying degrees of geometric complexity and the present work sets out to quantity the effect of varying geometric complexity on the agreement between industrial standard CFD simulations and experimental results. The experimental basis for the study is the work by Cermak [Physical modelling of the atmospheric boundary layer in long boundary-layer wind tunnels, in: Proc. Int. Workshop on Wind Tunnel Criteria and Techniques in Civil Engineering Applications, Gaithersburg, Md, USA, 1982] investigating the development of a ‘falsely’ thickened atmospheric boundary layer model. The upstream data are utilized as inlet boundary conditions to the computational simulation and results of the simulation compared to experiment at the downstream boundary. The geometric complexity takes the form of the thickening devices which are further upstream again from the inlet boundary. The necessity to manipulate experimental data to prescribe the required inlet boundary conditions to the flow fields are discussed. It is shown that simulations with the standard k- ϵ turbulence model fail to accurately model either the downstream velocity or turbulence intensity profiles. This discrepancy is attributed to the inlet turbulence structure differing from that inherent to the shear velocity profile. This discrepancy increases with the increasing complexity of the flow thickening devices.