**5. Results**

This considered case was a still rotor case. The tip speed ratio—tip speed divided by wind speed—was equal to zero because the rotor was fixed for each simulated angle.

Static torque—starting torque—is the torque required for starting turbine rotation. Figure 17 shows the relation between static torque and the rotation angle for the IceWind and Savonius turbines at an air velocity of 15.8 m/s. The static torque of the IceWind turbine was found to have two peaks at θ = 60◦ and 240◦. The Savonius turbine also showed two peaks but not at the same angles. The IceWind and Savonius turbines reached maximum values of 0.055 and 0.052 N·m, respectively. These slight differences may be because the flow field is two-dimensional near the Savonius rotor, whereas near IceWind rotor, it is three-dimensional. This fact will be proved later in the present study. It was found that the torque performance is improved by the IceWind rotor shape.

**Figure 17.** Relation between static torque and the rotation angle for the IceWind and Savonius turbines at an air velocity of 15.8 m/s.

Determination of the air flow velocity distribution and streamlines and pressure distribution around the turbine's surface enabled the air flow characteristics, disturbances, and locations of the highest pressure to be found. The air velocity distribution and streamline and pressure distribution results for both turbines at a wind velocity of 15.8 m/s are shown in Figures 18–20. The rotors of the two turbines showed similar flow patterns. The flow structure around the Savonius rotor was called "Coanda-like flow" by [21,22], and it controls flow separation on the convex side. A low pressure region forms on the side of the proceeding blade, contributing to the torque generation of the rotating rotor [23].

Figure 18a,b show the air flow velocity distribution around the still IceWind and Savonius turbines' rotors. These results were obtained at the plane that goes through the bottom of the turbines' blades as both have the same complete shape at this plane. The figures show similar velocity distributions. At θ = 0◦, the high fluid flow velocity moves at a tangent to the convex sides, while two circulating low velocity zones in the two concave sides are established. The air flow velocity has a slightly different maximum velocity of 28 m/s for the IceWind turbine compared with 22 m/s for the Savonius turbine. At θ = 90◦, the largest dead area is observed in the wake of the returning blade. The figures show similar velocity distributions. High fluid flow velocity touches the ends of both blades. Furthermore, a pair of asymmetric vortices develops behind both turbines. The smallest dead area is observed in the wake of the returning blade at θ = 0◦. Moreover, a maximum velocity of 34 m/s is observed for the IceWind turbine at θ = 30◦.

Figure 19a,b show air flow velocity streamlines around the still IceWind and Savonius turbines rotors at a velocity of 15.8 m/s. It is obvious from the figures that vortices behind the Savonius rotor are located between two imaginary planes that go through the top and bottom of the turbine blades. However, the rotor can be considered to be two-dimensional along the whole height. According to the top curvature of IceWind turbine, the plane that goes through the top of the turbine blades does not exist anymore. However, the plane that goes through the bottom of the turbine blades still exists. Vortices are located between the plane that goes through the bottom of the turbine blades and another inclined plane that follows the turbine top curvature. This provides IceWind turbine vortices with three-dimensionality. Vortices behind the IceWind turbine rotor appear to be larger than those behind the Savonius turbine.
