Introduction
Savonius-type wind turbines are vertical-axis "drag-type" machines characterized by high self-starting torque and omnidirectional operation. These traits make them particularly suitable for urban environments where wind is often turbulent and low-speed. However, their primary limitation is a low power coefficient ($C_p$). Optimizing blade geometry is essential for commercial viability, but the complex, unsteady flow around these turbines makes traditional experimental iteration costly. This study investigates the efficacy of 2D Computational Fluid Dynamics (CFD) using a "GTT code" as a predictive tool for design optimization, validated against Particle Image Velocimetry (PIV) data.
Methodology
The research utilized a dual-track validation approach:
- Experimental Visualization: PIV measurements were conducted in a water tank using a small-scale model to capture real-time velocity vectors.
- Numerical Simulation: 2D CFD simulations were performed using the GTT code to replicate the experimental flow fields.
- Performance Characterization: Wind tunnel tests measured torque and $C_p$ across various configurations, focusing on the Overlap Ratio ($OR$) and Blade Aspect Ratio ($AR$).
Results and Discussion
Flow Field Validation
The qualitative agreement between the PIV measurements and CFD simulations was high. As shown in Figure 1, the velocity vectors in the critical regions—specifically the flow entering the concave side of the advancing blade and the passage through the overlap—were accurately captured by the numerical model.
Optimization of Overlap Ratio
The overlap ratio ($OR$) significantly influences the turbine's performance by allowing air to pass from the returning blade to the advancing blade, thereby boosting torque. Both experimental and numerical results identified an optimal overlap ratio of $OR = 0.2$.
While the 2D CFD simulations overpredicted the absolute $C_p$ values—likely due to the absence of 3D "end effects" in the model—the trends perfectly matched the experimental data. As the overlap increased beyond 0.2 to 0.4 and 0.6, performance degraded significantly in both datasets.
Blade Shape and Aspect Ratio
Further investigations into blade geometry revealed that the standard circular arc ($AR = 1.0$) outperformed modified profiles. Increasing the aspect ratio (stretching the blade) to 2.0 resulted in a marked decrease in efficiency, suggesting that the traditional semi-circular profile remains highly effective for these drag-type rotors.
Conclusion
This study confirms that 2D CFD analysis using the GTT code is a reliable "directional" tool for Savonius turbine design. Although it lacks the absolute accuracy of 3D modeling, its ability to correctly predict performance trends across different geometries allows for rapid and cost-effective design iteration. The findings highlight $OR = 0.2$ and a circular arc profile as optimal parameters for maximizing power output.
Summary updated