OpenFOAM and its application to wind flow prediction

Abstract

The study of wind flow over complex terrain is a very important aspect of the study of pollutant dispersion. There are various methods in which this can be carried out. However, with the increasing availability of computational power the use of Computational Fluid Dynamics (CFD) modelling is receiving increasing attention. Solving the governing equations directly is often intractable and thus turbulence models are used. Hence, before CFD can be employed as a predictive tool, it must be validated against known or experimental results. There exist a large number of multi-purpose CFD codes. One of the recent additions to this array of tools is OpenFOAM. This software has attracted the attention of many users both - commercial and academic - because of its rapid development, extensibility, range of pre- and post-processing utilities, and most notably, because it has been made available as open source. However, given that it is still a rather new tool, OpenFOAM still needs extensive validation. The initial objective of this work was to validate OpenFOAM for use in predicting wind flows over a number of different obstacles. This was carried out by simulating a number of situations for which high quality wind tunnel experimental values were available. It was shown that the different OpenFOAM turbulence models can have different accuracies with the linear k-epsilon models attaining the best performance. Hence, one linear turbulent model was employed to predict the changes in wind flow pattern when balconies or non homogenous roof tops are present in urban street canyons.