A genetic algorithm approach to design aerofoils for small wind turbines

Abstract

A genetic algorithm (GA) methodology to design better site specific or wind profile specific aerofoils for small wind turbines starting from an initial pool of aerofoils is described. The GA methodology generates new aerofoils and quickly evaluates the fitness function of every aerofoil using XFOIL software. While XFOIL has a quick turnaround time, Computational Fluid Dynamics (CFD) methods are more expensive but have less assumptions. Hence, the suggested GA methodology can be used to quickly produce reasonably good test candidates for shortlisting the aerofoils for computational study. The best four aerofoils before GA and the best four after GA are taken and the top four performers from this pool are identified using CFD results. The results of the CFD study show that while three of the aerofoils after GA have better fitness values than the original pool, one from the original pool is still a comparatively good performer.A genetic algorithm (GA) methodology to design better site specific or wind profile specific aerofoils for small wind turbines starting from an initial pool of aerofoils is described. The GA methodology generates new aerofoils and quickly evaluates the fitness function of every aerofoil using XFOIL software. While XFOIL has a quick turnaround time, Computational Fluid Dynamics (CFD) methods are more expensive but have less assumptions. Hence, the suggested GA methodology can be used to quickly produce reasonably good test candidates for shortlisting the aerofoils for computational study. The best four aerofoils before GA and the best four after GA are taken and the top four performers from this pool are identified using CFD results. The results of the CFD study show that while three of the aerofoils after GA have better fitness values than the original pool, one from the original pool is still a comparatively good performer.