The aims of the research work described in this paper is a part were to use computational fluid dynamics (CFD) to investigate the factors affecting the performance of a single-stage downdraught evaporative cooling device for low-energy cooling of buildings developed from a novel prototype device described by Pearlmutter et al. (1996; 2008); and to model and explore the performance of the device when integrated within a hypothetical building. This involved carrying out simulations: to select the most effective wind catcher geometry. Two types of wind catcher using curved deflector and closed cowl design were studied: In total five alternative arrangements were investigated. Arrangements 1 and 2 were bi-directional wind catchers. Arrangement 1 was modelled without a baffle and arrangement 2 was modelled with an extended baffle. Arrangements 3, 4 and 5 were uni-directional closed cowls. Arrangement 3 was modelled without a baffle, arrangement 4 was modelled with a short baffle and arrangement 5 was modelled with an extended baffle and an increased inner radius of 1 metre which had the effect of raising the mid-plane height of the cowl inlet by 1 metre. Initially, for comparison in all studies, the inlet wind speed was set at 10 m/s at a reference height of 11.5 metres which corresponded to the mid plane height of the wind catcher and wind cowl entry ducts for arrangements 1 to 4. The CFD simulations were carried out to define the optimum geometry of a wind catcher.