Taguchi method-based performance evaluation of a Savonius turbine augmented with winglet deflectors

Deflectors have been shown to improve the power output of Savonius turbines. However, their effectiveness is limited by the formation of tip vortices at the edges of the deflector—particularly at the returning blade deflector oriented normally to the oncoming flow. These vortices induce flow instabilities that degrade turbine performance. To mitigate this issue, a winglet was integrated into the returning blade deflector and its effects on turbine performance was investigated through wind tunnel experiment. The study examined three winglet parameters, each varied at two levels: winglet length (10 mm and 30 mm), cant angle (0° and 45°), and orientation (forward [+] and backward [–]). Based on the Taguchi method, the first 4 runs of the L₈(2³) orthogonal array were analyzed to assess the influence of winglet parameters on turbine performance and to predict power performance. The analysis identified winglet orientation as the most influential parameter, followed by cant angle and winglet length. Moreover, the Taguchi method proved effective in capturing these effects, accurately predicting the maximum coefficient of power in 6 out of 8 runs. The optimal winglet deflector, i.e. 30 mm length, 45° cant angle, and [+] orientation achieved the highest power coefficient of 0.257, representing a 14.5% improvement over the baseline deflector. Flow visualization further confirmed that this design weakened tip vortex strength and shortened the moment arm on the returning blade, thereby enhancing turbine performance.