**1. Introduction**

In general terms, the market for small wind turbines is currently growing, although the sector of small wind turbines intended for installation in buildings is increasing at a lower rate. According to the World Wind Energy Association (WWEA) [1], the installation of small wind turbines will increase by around 12% annually in the 2015–2020 period. The good economic profitability of small wind turbines and the consistency of technological advancement are determinant factors that explain the growth of the small wind turbine market. On the other hand, in the last several years, research is increasingly being focused on the development of different technologies that help minimize the energy consumption of buildings. This philosophy is known as nZEB (nearly Zero-Energy Building) [2], and it is included in the EU 2010/31/CE directive related to the energy efficiency of buildings. After 2018, every new public building should be constructed in accord with this regulation and, after 2020, every new building should be compliant.

The goal is to maximize energy efficiency and reduce the use of primary energy derived from fossil resources so that the required energy demand can be met by renewable sources. In this sense, mini wind technology, which involves generating energy with wind turbines of 100 kW or less to cover an area smaller than 200 m2, can play a very important role. However, some technological challenges, such as the vibrations, the generated noise levels, and the device's aesthetic and architectonic integration, are ye<sup>t</sup> to be fully solved.

Nevertheless, these devices have many advantages:


The recent developments in wind energy for urban environments have inspired different types of Building-Integrated Wind Turbine (BIWT) projects. For example, in London, Strata SE1 is a tall building with 43 floors that will include three wind turbines with diameters of 9 m on the roof of the structure. These wind turbines will be used to meet the building's lighting demand [3].

On a smaller scale, there are a lot of projects that include Horizontal-Axis Wind Turbines (HAWTs) integrated with buildings, as well as Vertical-Axis Wind Turbines (VAWTs). These projects are focused on integrating wind turbines with existing buildings. Thus, these buildings were not previously designed to accelerate air streams, unlike the World Trade Center of Baharein [4] or the mentioned Strata **SE1** building. According to this post-integration trend, building-integrated wind turbines are being implemented in strategic locations to capture the acceleration of air streams that are produced because of different geometries. In this sense, the most interesting locations are the upper and lateral edges of a building, especially the former because it is at a reasonable distance from homes.

Nowadays, there are several ongoing projects working to develop an optimal system that harnesses wind energy in urban environments. Most of them have concluded that wind turbines located in obstacle-free environments are not adequate for urban environments because of the urban turbulent flow, which can present a relevant turbulence intensity on the superior edges of the buildings [5,6]. For that reason, HAWT devices, which usually exhibit good performances with laminar flows, perform poorly in urban environments, in addition to their generation of noise as high as 200 dB within a radius of 500 m [7]. Conversely, VAWTs play an essential role in generating wind energy in urban areas since their performance is not much affected by turbulent flows, and they tend to be noiseless [8]. Additionally, the VAWT has a lower cut-in speed than HAWT and a larger or even unlimited cut-off speed, ensuring longer operating times [9–11]. Although the power coefficient is lower, the design is simpler and the manufacturing process is easier to carry out.

Along these lines, the existing urban wind energy potential has encouraged researchers to develop a proper methodology for wind energy estimation in urban environments [12]. The use of anemometers at specific locations can be combined with advanced computational simulations of buildings situated in complex urban terrains using CFD (Computational Fluid Dynamics). In this way, wind energy potential estimation using reanalysis and meteorological mesoscale models, which is a well-known offshore and onshore method and also developed by the authors [13,14], can be complemented with different back-end tools.

In this work, the authors present the design of a Savonius drag-driven turbine that is intended for integration into buildings. The proposed turbine is called ROSEO-BIWT, which has been specially designed to work in urban environments. The wind in urban areas is characterized by its turbulence, thus it is important to take advantage of low-speed air streams. The germinal project of ROSEO won the first award in the EDP-RENEWABLE UNIVERSITY CHALLENGE 2017 [15], and the members of the project have now created a university start-up called ROSEO. Although it is typically used as a vertical-axis turbine, ROSEO-BIWT is formed by a Savonius turbine in a horizontal position and concentration vanes that accelerate the air streams by the Venturi effect (see Section 2.2). These types of vanes are usually called PAGVs (Power Augmentation Guiding Vanes) [16–18]. The proposed turbine was also designed to be easily architectonically integrated. This was the case for the design proposed by Park et al. [19], in which several Savonius turbines were incorporated into the facade of a building at different heights to take advantage of the vertical currents created by the wind on the walls of the building.

The Savonius wind turbine is a drag-based device, unlike the majority of turbines, which are lift-based. This particular aspect allows for low noise levels and few vibrations, and these factors are very important in building installations [20,21]. The PAGV increases the wind speed as the catching area grows, resulting in a system that is able to start at wind speeds of about 1 m/s, thus ensuring a grea<sup>t</sup> number of energy-producing hours. Furthermore, energy generation continues no matter how high the wind speed is.

This paper proceeds as follows: a possible location for the installation, which was established using ERA5 data, is presented. ERA5 is a powerful tool for global atmospheric analysis that is updated in real time (see Section 2.1). The authors also installed an anemometer on the roof of their university to calibrate the wind data for a period of eight months against ERA5 (Section 2.1.2). In this way, an empirical method for the estimation of wind energy potential on buildings with a low computational cost will be developed in subsequent work, as discussed in Section 3.5. Finally, a preliminary small-scale experiment was developed for a wind tunnel with a small Savonius and different configurations of the PAGV (Sections 2.2 and 2.3). The authors finish this work with some relevant conclusions and a future outlook of possible research directions. The qualitative methodology used here can be considered within the scope of analogical reasoning and model construction [22].
