Vortex Generators Project for an Unmanned Small Airplane

João Barreto Duarte Neto joaobarretomecufc@gmail.com orcid.org/0000-0002-7979-0871 Universidade Federal do Ceará (UFC), Fortaleza, Ceará, Brasil. Maria Elisa Marciano Martinez melisa@inpi.gov.br orcid.org/0000-0002-8010-869X Instituto Nacional da Propriedade Industrial (INPI), São Paulo, São Paulo, Brasil. Marcello Carvalho dos Reis marcello@meteora.com.br orcid.org/0000-0002-1132-9034 Meteora, Fortaleza, Ceará, Brasil. Claus Franz Wehmann claus.wehmann@ufc.br orcid.org/0000-0001-8756-9387 Universidade Federal do Ceará (UFC), Fortaleza, Ceará, Brasil. Since the Cold War period, turbine generators have proved to be an important alternative to the development of aerodynamic aircraft designs, and even so, there was little adherence to their use, among which we highlight, short takeoff and landing aircraft (STOL) and some models of military aircraft and commercial jet aircraft. In the USA and Brazil, in addition to other countries, they hold competitions to awaken technological innovation in the aeronautical field for engineering students: the SAE AeroDesign Competition (SAE Society of Automotive Engineers). These contests allow teams from their countries to use points of difference in their design and, in some cases, to use small unmanned aircraft devices that have already been designed for huge aircraft, such as vortex generators. Therefore, we intend to adopt a Vortex Generator model for the Avoante Aeromec AeroDesign team airplane project (team from the Federal University of Ceara, Brazil UFC), use Computational Fluid Dynamics (CFD) with turbulence modeling, as well as perform tests in a wind tunnel. Despite the problems found in comparing the results of the computational models with the prototype tests, it was possible to prove the efficiency of these vortex generators in the design of the team's airplane, observing the increase in the stall angle, reduction of the drag coefficient and increase of the coefficient lifting.


INTRODUCTION
In aircraft designs, we always take into account how the air behaves when the aircraft is subjected to different conditions. In the case of aircraft designs, these conditions can vary considerably, given that all the variables that make the flight process possible can be changed drastically, and if the project does not provide for them, the final product -the aircraft or plane -will have a high risk of failure, that is, an accident. For example, an airplane must be prepared for the takeoff, cruise and landing phases, although the potential gust and rain conditions must also be considered because they are extreme with regard to the aerodynamics and stability of the project (ANDERSON Jr., 2001).
In the USA and Brazil, in addition to other countries, they hold competitions to awaken technological innovation in the aeronautical field for engineering students: the SAE AeroDesign Competition (SAE -Society of Automotive Engineers). These contests allow teams from their countries to use points of difference in their design and, in some cases, to use small unmanned aircraft devices that have already been designed for huge aircraft, such as vortex generators.
In aircraft designs, the use of devices that improve the aerodynamic efficiency and stability of the aircraft are considered, above all, the aerodynamic part must observe the air flow above and below its wing. This air that is closest to the wing surfaces is not moving evenly, this is a thin layer to which the fluid's viscosity affects its flow, known as the boundary layer. This layer influences the variation of force acting on the lifting surface, especially changing the stall and drag properties. Generally, he tried to prevent it from flowing off the surface, because that would greatly change the pressure distribution configuration thought of in the wing design; this new distribution can considerably affect the aircraft's flight and its external structure (GROSS;FASEL, 2018;LAMBERT;RAZAK;DIMITRIADIS, 2017;DURDEN, 2014).
Based on these observations, the vortex generators were designed as small devices that would prevent stall and turbulence, increasing the interaction between the fluid layers that flow alongside the wing surface, which maintain this flow in this region even at high angles attack (AoA). Vortex generators (VGs) are devices placed on the elevation surfaces of an aircraft (DURDEN, 2014;CALLE, 2015) -especially on the wing -to prevent the displacement of the boundary layer and, thus, increase the stall angle, improve the stability and shorten landing and takeoff runway lengths. They work by creating a small vortex that generates a secondary flow that avoids the loss of moment of the flow and its displacement from the wing surface (DUARTE NETO, 2019). These vortices will cause the air flow -more specifically the boundary layer -to mix and remain close to the upper wing surface, even at high angles of attack, thus increasing the stall angle. This behavior is used to decrease drag and increase the support of aerodynamic structures in several areas, for example, wind energy (BALDACCHINO et al., 2018), marine engineering (AHMED, ELBATRAN; SHABARA, 2014) and the automotive industry (LÖGDBERG, 2006).
The ability of VGs to reduce drag is of interest to small planes such as unmanned aerial vehicles (UAVs) due to the limitations inherent in these devices, such as engine power, battery, fuel tank size, weight, among other variables. Therefore, some dedicated solutions are developed using passive devices such as VGs (ZHEN; ZUBAIR; AHMAD, 2011;MARQUES;BACHOUCHE;MALIGNO, 2013;PAIBOOLSIRICHIT, 2016;SATTAROV, 2018;SATTAROV et al., 2019). In the case of competition, such as SAE Aerodesign, much effort is devoted to the development of radio controlled airplanes for the best performance (MIYADAIRA; CASTRO NETO; CARVALHO, 2018;COSTA et al., 2018;LEHMKUHL;MARIGA;OLIVEIRA Jr., 2018;REIS et al., 2018), with the objectives being short takeoff length, landing and takeoff weight (WTO). There are different strategies used by the teams, some of them use the reduction of induced drag (REIS et al., 2018;TURCATO et al., 2018) and, more recently, the suggested VGs (MELO; MARIANO, 2019). The main problem addressed in this work is to adopt a VG design for the Avoante Aeromec AeroDesign team's airplane design, using turbulence models in CFD and wind tunnel tests, since this use of technology is a matter of commitment and, as pointed out by Baldacchino et al. (2018), not always effective.

MATERIALS AND METHODS
In the prototypes made, the CAD software was used for drawing the drawings and, for the analysis, the CFD software. All dimensions shown in these two software are in the millimeter scale.

PROJECT
To compare the results found in this study with the results of (MELO et al., 2019), it was used both in the analysis of the CFD and test in a wind tunnel whose wing had a wingspan of 400 mm whose rope with the value of the aerodynamic mean chord -or CMac -, that is, 382.22 mm, and the control volumes were designed with this chord value using the Mullen (2019)   As the objective was to verify the difference of the VGs in relation specifically to the airfoil, it compares the 2D calculations of the airfoil without VGs with the three-dimensional (3D) analysis of the airfoil with VGs. This method was worked on in Sørensen, Zahle and Vronsky (2014), which uses an opposite rotation configuration and explores geometric symmetry by simulating only one of the reeds. In contrast, this work simulated an entire group of VGs for a complete wingspan.
The 3D is shown in Figure 2 (with the same dimensions as the 2D, but with a width of 400 mm -the same measurement of the wingspan), because the purpose of this analysis was to compare these results with the results of the tunnel wind). Aerodynamic analyzes used the turbulence model known as Spalart-Allmaras, as suggested by Cornell University (2015), which was designed and optimized for wing and airfoil flows and produced favorable results without the gradation of the use of computational tools. These analyzes were performed for four different speed values (7.33 m/s, 10.45 m/s, 15.32 m/s and 16 m/s), and for each speed the angles from 0° to 25° (ranging from 5° to each other), in addition to the stall angle, incidence angle, climbing angle and minimum angle described in .
For the analyzes with the GPs, the model used by Sørensen, Zahle and Vronsky (2014) was chosen, with the format of Figure 3 and the dimensions of  The execution model chosen was the most relevant, as pointed out by Fouatih et al. (2016) and Kumar et al. (2016), given that the counter-rotating device appears as the most effective agreement with Godard and Stanislas (2006), and was also used in this study. The positioning configuration for the entire wing extension follows in Figure  5, with the distances proportional to the airfoil chord value; as in Sørensen, Zahle and Vronsky (2014): *…+ the height (h) of the VG is 1 percent of the chord length, the aspect ratio (l/h) of the VG is approximately 3.8 and the inclination with the incoming flow (β) is 15.5 degrees, the distance between the VG within the pairs is Δ1/h ∼ 5 at the leading edge of the VG, while the distance at the leading edge to the next VG pair is Δ2/h ∼ 4.

PROTOTYPES
The wing was the same used by , just with some repairs, as it is introduced in Figure 6. The manufacture of the VGs was made in a 3D printer using ABS (acrylonitrile butadiene styrene -plastic material) with later adjustment using sandpaper to remove the imperfections, as seen in   The wing remained attached to the upper end of the scales (Figure 9); in this case, there were two sensors that read the deformations due to lift and drag on the wing. Speed was measured by an anemometer (shown on the left side of Figure 9).
The tests were done with the same angles of the analyzes and speeds at which the tests were carried out were 7.33 m/s and 10.45 m/s, the first being the minimum speed of the unmanned airplane and the second is the takeoff speed.  Figures 10 to 13 follow with their labels:    The results were consistent with those proposed by Sørensen, Zahle and Vronsky (2014) and Seshagiri, Cooper and TRAUB (2009) given that for the speeds used, the VGs made the drag decrease and the lift increase, with an increase of 30% in the CFD model and 12,5% under test in the wind tunnel. These results are in agreement with Seshagiri, Cooper and TRAUB (2009) if the differences in angles used in both studies are considered.

There were made the graphs C L x A o A and C D x A o A to each one of the velocities in the CFD's analyses and in the wind tunnel tests. The
The drag reduction was measured in the CFD model in the last 40%, and the same was not observed in the wind tunnel test, since the drag in the wind tunnel test includes induced drag and in the CFD, it does not. In the analysis carried out, a 2D model was used, despite having the same shape transversely, in wingspan it is different from the real wing of the project, which contradicts the similarity of flow presented by Anderson Jr. (2001) and Fox, McDonald and Pritchard (2011) and, as expected, the lifting is greater and the drag less, as shown by Anderson Jr. (2001) and Calle (2015). In the end, it was identified that the same in all tested speeds.
With the CFD data, it was possible to observe that the greater the angle, the greater the difference made by the VGs, at least up to the stall angle, this was observed in different types of VGs, as demonstrated by Kerho et al. (1993), and, Barrett and Farokhif (1996), and with similar effectiveness. This behavior was also demonstrated in a wind tunnel test, with maxima close to the stall angle with a difference of 12.5% for a lower speed and 16% close to 10° for a speed of 10.45 m/s. Both situations are essential for the developed prototype, because in addition to avoiding the stall and it actually occurs at the angle designed for takeoff.

CONCLUSIONS
With this study and prototype development, it was possible to compare the wings of unmanned airplanes with and without vortex generators, both in CFD and in wind tunnel tests, even with the use of 2D analysis and the nonconsideration of drag from the tip of the wing, it was noted that the goal was achieved with the VGs.
It was also observed that the VGs increased the stall angle of the airfoil for all cases, as in the CFD and in the wind tunnel test, increased the lift, especially at greater angles, as well as reduced drag. The CFD analysis showed that the use of the VG could increase the increase in the last 30% and reduce drag in the last 40% in the best scenario. The wind tunnel test finally showed a 16% better lift, even if the reduced drag was negligible.
The same behavior had already been observed by other authors (SESHAGIRI;COOPER;TRAUB, 2009;SØRENSEN;ZAHLE;VRONSKY, 2014;ZHEN;ZUBAIR;AHMAD, 2011;PAIBOOLSIRICHIT, 2016). These results confirm that the use of the methodology described by Sørensen, Zahle and Vronsky (2014), even if it was not developed for small-scale aircraft like UAV, fits the proposed improvement of aerodynamic characteristics and performance in the developed prototype. The difference in 2D CFD and Experimental data in wind tunnel test t results from limitations as it is a 2D flow analysis in computational analysis. As expected, the CL and the CD are the same for an airfoil and the wind tunnel test was carried out with a small wing, which, however good the approach is. It is worth mentioning that the wind tunnel tests showed drag values considerably higher than those of the CFD, since the drag of the wing tip (induced drag) is dominant in the results in the wind tunnel test.