Degree Name

Masters by Research


School of Mechanical, Materials and Mechatronics Engineering


Wind turbines can be classified into different categories based on the power output: large (>1 MW), medium (40 kW–1 MW) and small (

On-site power generation from micro wind turbines properly located on roof - tops can supply carbon free energy whenever it is needed, thus decreasing our dependence on fossil fuel, and helping to reduce carbon dioxide (CO2) emissions. However, the viability of micro wind turbines as a potential power source for domestic energy needs is still hindered by i) the unfavourable characteristics of urban wind such as high turbulent flow and low velocity, and ii) the limited output of micro wind turbines.

Since the performance of micro wind turbines is influenced by the wind characteristics of a suburban topology, it is important to study the location and conditions where micro wind turbines should best be placed.

An increase in wind speed across the turbine rotor significantly increases the power output, so encasing micro wind turbines inside ducts increases the mass flow rate drawn into the turbine which generates an output power typically 2 times higher than conventional wind turbines. Moreover, the wind speed can be further enhanced by adjusting the duct geometry with diffuser shaped ducts. Diffuser-shaped ducts have been experimentally proven to increase the performance of micro wind turbines. Despite this, only limited computational work with simplified turbine geometries has been undertaken.

The present work used the Computational Fluid Dynamics (CFD) technique to investigate various factors affecting the performance of diffuser augmented micro wind turbines. Firstly, the characteristics of wind flow above the roofs of typical suburban buildings (pitched, pyramidal, and flat roofs) were investigated to find the optimum roof top position for a wind turbine. Secondly, the effect of various diffuser geometries on the efficiency of micro wind turbines was studied. The coefficient of performance of a straight diffuser, an airfoil shaped diffuser, a diffuser with flanges, an airfoil shaped diffuser with flanges, and a compact brim diffuser were compared to conventional bare wind turbines. The flow field around the turbine blades and inside the diffuser was analysed in terms of developed 3-D stream lines, the pressure field, velocity field, and mass flow rate. Lastly, the influence of rotor features such as the profile of the rotor blade, the number of blades, and the clearance between the blade tip and the duct, on the output power were evaluated. It should be noted that the applied methodology was validated with the numerical and experimental results available in the literature.

The results confirmed that the wind flow characteristics are strongly dependent on the profile of the roofs as well as the direction of the incoming wind. The wind above flat roofs had not only lower turbulent intensity compared to the other roof profiles, the turbulence beyond the roof also decreased rapidly. In terms of wind velocity, both pitched and pyramidal roofs generally slow the wind down. Therefore, turbines located on flat roofed houses are likely to yield higher and more consistent power for the same turbine hub elevation than the other roof profiles.

All the diffuser configurations that were investigated showed at least 25% higher mass flow rate through the turbine rotor compared to a bare wind turbine. It was noted that the radial velocity was larger on the tips of the bare wind turbine blades than the diffuser turbines, hence the tip vortex on the bare wind turbine was stronger leading to a lower efficiency. Of all the diffusers studied, the compact brim diffuser had the highest coefficient of performance, not only was there a significant speed up effect, there was also a considerable drop in pressure across the turbine.

The performance of the diffuser augmented micro wind turbines (DAWT) at high tip speed ratio (λ) values depended primarily on the rotor features, but only minimally on the blade profiles at low to medium λ range (2.5 < λ < 3.5). Furthermore, the existence of an optimal tip clearance is revealed.

FoR codes (2008)




Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.