The development and experimental validation of a simulation method for the optimisation of diffuser augmented wind turbines

The addition of a diffuser to a small Horizontal Axis Wind Turbine (HAWT) can significantly improve the performance of the wind turbine. It has long been posited that the resultant wind turbine, known as a Diffuser Augmented Wind Turbine (DAWT), can have improved performance over a bare turbine of the same size with the additional benefits of improved noise mitigation and operation in turbulent wind conditions. It has been shown that the performance of the diffuser affects the performance of the wind turbine blades and vice versa, however, previous work in the field of DAWT optimisation has focussed either on optimising blades for a specific diffuser, or optimising a diffuser and adding a standard HAWT for use within it, both of which are not ideal for the optimisation of the complete system. The work presented in this thesis documents the development and implementation of a new simulation methodology, designated the Combined CFD/BEM Method, to predict the starting performance and power output of a DAWT. This new approach allows the diffuser to be optimised in conjunction with the turbine blades, a process that is computationally intensive using existing methods. This new method can rapidly assess the performance of the DAWT, using optimisation processes heretofore too computationally expensive to be practicable. Effective optimisation techniques which require thousands of geometries to be assessed, such as genetic algorithms, can be implemented using the Combined CFD/BEM Method to optimise the geometry of a DAWT. The optimisation process described in this thesis has utilised genetic algorithms to maximise the Annual Energy Production (AEP) of a DAWT located at a typical small wind turbine site. The wind data for the typical site was obtained from the International Electrotechnical Commission’s (IEC) standards for small wind turbines. Using rapid prototyping methods, the two highest performing DAWTs nominated by the Combined CFD/BEM Method, with diffuser radii of 0.203 m and 0.217 m, were constructed and experimentally tested in the large recirculating wind tunnel at the University of Newcastle, Australia. Two HAWTs with blades of the same radii as the diffusers of the two DAWTs were also optimised and constructed for experimental testing to compare to the performance of the DAWTs. The different wind turbine configurations were experimentally tested for starting performance and mechanical power output for wind speeds between 5 m/s and 10 m/s inclusive. The Combined CFD/BEM Method was shown to accurately predict the starting wind speed of the tested geometries, which were all within 11.4% of the experimental results. The combined CFD/BEM Method underpredicted the power output at the lower tested wind speeds and overpredicted at the higher speeds. For the lower wind speeds, the largest underprediction was 26.5% at 5 m/s with the majority of the predicted power outputs within 10% of experimental results. The largest overprediction at the higher wind speeds was 9.3% at 10 m/s. The Combined CFD/BEM Method has also shown that a DAWT can exceed the power output and AEP of a HAWT of the same size although, given the material utilised for rapid prototyping, this could not be fully validated as the printed blades lacked sufficient rigidity for comprehensive testing. It is worth noting that the diffuser geometries assessed for the current work were simple revolved aerofoils, unadorned with vortex generators, gurney flaps, boundary layer control slots or any other devices that have been shown to improve the performance of a diffuser. Given the accuracy of the Combined CFD/BEM method, it could be utilised in future research to optimise a DAWT employing any of these performance-enhancing modifications.