2.2.3. The Final Design

Park et al. [19] developed the idea of using a larger facade surface to generate energy by installing a lot of Savonius rotors at different heights while also using PAGVs to improve the performance of the wind turbines. The system that they proposed is similar to a ventilated facade. It is important to emphasize that they wanted to capture the vertical air streams that are generated on the windward side of the building, as in our case. However, they used parallel vanes in the facade with a small concentration angle; in our case, the upper edge is used to augmen<sup>t</sup> the concentration angle and capture not only the vertical stream on the facade but also the horizontal component of the wind. Furthermore, the background of the Savonius rotor is free on the edge of the building: this is an important aspect that is not encountered in turbines installed in the facade. Although the influence of this aspect is out of the scope of this study, it is an obvious aerodynamic advantage.

Another innovation is that the proposed turbine can be installed in existing buildings: it is not a design intended only for new buildings. To summarize, ROSEO-BIWT shows good architectural integration in existing buildings and high economic viability due to the simplicity of the design.

Figure 4 shows the ROSEO-BIWT design. This is a schematic perspective that does not take into account the influence of the angle between the two PAGVs; the angle can be adapted for other positions of the vanes.

**Figure 4.** ROSEO-BIWT design.

The PAGV areal ratio between the entrance and exit of the air is 4:1, and a similar *AF* is expected to result from a first simplistic calculus due to the Venturi effect. In any case, as mentioned, the complex interaction between the outlet wind speed and the rotor motion deviates this a priori value of *AF* = 4.

In their seminal work about a curtain design to increase the performance of a Savonius turbine, Altan et al. [46] determined that the best angles for the capture of wind in their curtain design are 15◦ for the superior vane and 45◦ for the inferior one that obstructs the negative torque. The authors corroborated the same influence of the inferior vane in the laboratory (see Section 3.1). Figure 5 shows the dimensions for this optimum design with curtains. It is considered a unit of capture width at the inlet, and a geometrical relation of 4:1 for the inlet width (0.25, therefore diameter of 0.50) versus the outlet width. The aspect ratio is six considering the results of Roy and Saha [45]: 3 = 0.5 × 6. These proportions can be established between a capture width of one and two meter; within this size, ROSEO device is manipulable for a worker on the roof in the implementation process and for O&M issues, without the need of a crane .

**Figure 5.** Adapted figure of the optimum design using curtain vanes [46].

#### *2.3. Experiments in the Wind Tunnel*

Although there are results provided by previous studies, in the following sections, the authors describe the general experimental methodology that is being developed. The experimental model construction is proposed by referencing previous findings and design procedures [22]. These are the main steps:


Table 1 describes the main characteristics of the above-mentioned wind tunnel. Figure 6 shows the wind tunnel and the installation of a PAGV and a real Savonius rotor. It should be mentioned that the disposition of the vane below the limit of the rotor's horizontal axis obstructs the negative torque and, simultaneously, accelerates the stream in the opened drag side above the axis. As mentioned, this aerodynamic effect has been properly documented in previous reviews about the performance of the Savonius rotor [39,41,47].


**Table 1.** Characteristics of the wind tunnel.

The augmentation may be even higher because of the corner effect of our design. However, until now, these preliminary measurements have only been performed with low values of steady wind speed without considering some important effects, such as the blockage ratio of the tunnel [49,50]. However, the the influence of *AF* is important at these low wind speeds below the rated power, because it can ensure a sufficient wind speed above the rated wind speed at the outlet of the concentration vanes.

On the other hand, Figure 7 shows the small-building, the guiding vanes with 3:1 inlet/outlet relation, and the six-bladed rotor of 2 cm diameter. Here, the objective is to create an anemometer that is able to capture the wind on the entire outlet area of the vanes. According to our previous experiments, measurements with Pitot tubes result in grea<sup>t</sup> fluctuations due to small displacements or inclination deviation in such a narrow area. In this case, the electrical motor is a *maxon DCX06M EB KL 6V* of 0.529 W. The speed constant is of 3060 min−<sup>1</sup> V −1 having a direct way to compute the angular velocity in function of the voltage. Additionally, there is also a constant speed–torque relation of 36,600 min−<sup>1</sup> mNm−1. This characteristic is important because it allows computing the increment of the torque due to the increment of the speed.

**Figure 6.** The Savonius turbine inside the wind tunnel with the inferior vane.

**Figure 7.** The small building model (**a**) without the vanes; and (**b**) with the vanes.
