*WINNER* Simulation and Optimization of Wind Turbine Geometries

Abstract

In this study, the results from academic wind turbine simulation software were compared to viscous unsteady computational fluid dynamic simulations done on the Tennessee Tech High Performance Computing cluster (HPC). The cross section of the vertical axis turbine blade was altered by means of a genetic algorithm in order to produce the highest power coefficient. Python was implemented to automate the setup of the simulation, running the HPC, and post processing data to be inputted back into the algorithm. The simulations were done over a variety of rotational velocities in order to better match a real turbine.