Computational analysis of Savonius wind turbine modifications including novel scooplet-based design attained via smart numerical optimization

Abstract

Simple design and low power coefficient of Savonius-style wind turbines make them attractive in experimental and computational fluid dynamics (CFD) based optimization studies. Still, after numerous studies the power coefficient is low, and the single optimal Savonius-style wind turbine shape and configuration is not known. This leaves room for improving the power coefficient, and the objective of this paper is to find an improved solution with application of smart numerical optimization.Several Savonius blade improvements proposed by other researchers were initially analyzed. The numerical (3D CFD) analysis has shown that simple modifications of the blades can be better than more complicated shapes such as spline or elliptical curve. This has encouraged the investigation on what is the limit to the optimized design which used only simple circular-arc based blades. Thus, we perform a global optimization of Savonius blade pair where each pair is composed of two circular arc segments. The smart numerical optimization was performed using 2D CFD and the results were verified and compared with the previous designs using the 3D CFD. The selected optimization operating conditions were tip speed ratio 0.9 and wind speed 12 m/s. The novel optimized design achieved 39% improvement of the maximum power coefficient relative to the classical Savonius design.