Wind Tunnel Aerodynamic Study of a VTOL-UAV for Wide Range of Angles of Attack and Sideslip Angles

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Springer Nature in Aerotecnica Missili & Spazio, Vol. 100, on 2021, available at: https://doi.org/10.1016/j.automatica.2020.109474 © 2020 AIDAA Associazione Italiana di Aeronautica e Astronautica. Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles (EMERGENTIa project) Victoria M. Prieto ∗, Sergio Esteban∗∗ Higher Technical School of Engineering University of Seville Abstract This article presents the design, manufacturing and wind tunnel tests conducted on a 1:4 scaled Vertical Take Off and Landing (VTOL) Unmanned Aerial Vehicle (UAV). A briefly description of the VTOL model, identifying characteristic dimensions, along with the manufacturing process for the scale model and the measurement equipment is presented. Results of the wind tunnel experiments are conducted varying both angle of attack and sideslip angle for a wide rage of the angles, and several wind tunnel speeds, thus allowing to characterize the full flight envelope for all three flight phases of the VTOLUAV: axial, transition and cruise flight phases. A description of the wind tunnel measuring setup and the methodology used is presented, along with the subsequent analysis and corrections considered for all the data obtained, presenting a discussion about the results and its comparison with numerical analysis obtained with CFD studies. Finally, some conclusions and recommendations are made with the main objective of establishing a basis for future tests. Keywords Wind Tunnel ·Angle of attack ·Sideslip angle ·Solid Blockage ·Complete flight envelop ·CFD 1 Introduction This article presents wind tunnel experiments conducted on a convertible Vertical Take Off and Landing (VTOL) Unmanned Aerial Vehicle (UAV) in order to characterize the entire full aerodynamic flight envelope. The motivation behind these experiments come from the need of having precise aerodynamic models that will serve as bench mark where to test the innovative design presented within the EMERGENTIA project. The EMERGENTIa project (acronym for the full title: DevElopment or fan unManned convERtible aircraft for rapid and efficient deployment in emerGENcy situationTIons) arises from the fact that emerging technologies are radically changing traditional operating procedures in disaster relief and emergency response management. In particular UAVs are proving themselves extremely useful in Search And Rescue (SAR) missions. They bring very advantageous capabilities such as rapid response, remote operation, transportation of equipment, monitoring of wide areas, and multi-sensor deployment. Fig. 1 VTOL-CP1 prototype Rapid Intervention Vehicles (RIV) typically used in SAR missions have cargo space for at most one small Remotely Piloted ∗Victoria M. Prieto, Master’s Degree in Aeronautical Engineering, [email protected] ∗∗Sergio Esteban, Associate Professor, [email protected] Aerospace Engineering and Fluid Mechanics Department, ETSI Sevilla 2 Victoria M. Prieto, Sergio Esteban Aircraft System (RPAS) of either fixed or rotatory wing type. Fixed wing aircraft are well suited to rapid deployment and remote monitoring but require external means (runways or catapults) for takeoff and landing, and cannot hover. Rotatory wing aircraft (such as quadrotors) have much less autonomy and range but can perform Vertical Take Off and Landing and hover; thus they are suited for restricted or inaccessible areas and sensor deployment. A third kind of aircraft, still relatively uncommon, is fixed-wing convertible aircraft with VTOL capabilities. They have the advantages of fixed and rotary wing aircraft without most of their shortcomings. A convertible UAV would allow a first response RIV to operate a single multipurpose, multi-mission aircraft. Project EMERGENTIA proposes to explore the use of a convertible fixed-wing aircraft with VTOL capability specifically designed for SAR missions, combining the virtues of fixed wing and rotary wing aircraft by adopting the type of aircraft that is called to transform the way in which it operates in all sectors of society given that they have the advantage of being able to operate as well as a rotary wing aircraft (performing hoovering flight, vertical take-off and landing), as a fixed wing aircraft (gaining greater range and autonomy), being able to carry out the transition between both modes of operation in flight.The aircraft will be easy to operate and fast to deploy. To this mean, guidance, navigation and control algorithms will be developed taking into account the requirements of SAR missions, while ensuring safe communications and operations of the drone using advanced cyber-physical systems, since fully autonomous operations can be supervised only by a healthcare emergency professional without any specialized training. In order to achieve this objectives it is necessary to have robust control strategies that will allow the UAV to operate autonomously. In addition, in order to test the robustness of the different autonomous control strategies, high fidelity Six-Degree-of Freedom (6DOF) simulator are required. This high fidelity 6DOF simulator does require having complete aerodynamic and propulsive models for the entire flight envelop of the VTOL-UAV. In addition, in order to define the flight envelope, it is necessary to identify the different flight phases of a typical SAR mission: vertical takeoff from the advanced command post (axial flight phase), once reached the desired altitude the aircraft would transition to a fixed-wing aircraft flight mode (transition flight phase), conduct a flight mission (cruise flight phase) within the radius of action modeling a scenario of an inaccessible area due to a catastrophe carrying the prescribed payload (communications equipment, food, automatic defibrillator, or transport of blood derivatives of sensors for automatic triage) that must be delivered in the area affected by means vertical landing, with subsequent vertical takeoff, with the possibility of collecting samples (for example in contaminated areas), return to the command post, and conduct a vertical landing. The article here presented tries to characterize the full flight envelope aerodynamic model for all three flight phases by conducting wind tunnel experiments on the VTOL prototype designed for this project varying both angle of attack and sideslip angle, throughout a large range of values. Not too many wind tunnel experiments have been found in the literature for high angles of attack, within the works found, they are mostly limited to high angles of attack experiments for turbine blades [22,20]. Wind tunnel experiments of similar aircraft configurations are also found in [18], but limited to the study on the effect of tail dihedral angle on lateral directional stability due to sideslip angles. When referring to full aircraft wind tunnel experiments, the most common disposition is the one used in [12] that, although it allows combining up to three different angles at the same time (attack, sideslip and balance), it is clearly limited to small values of these angles. Since the main purpose of this paper is to characterize the full envelope of the aircraft being studied, it was decided to use a similar configuration to the one used in [22] and [20], even though some modifications had to be made in order to adapt it to the analysis of the full aircraft instead of a single wing. Also, other configurations were considered, as those shown in [11] and [10].Further specifications of the final configuration are given in section 2.1. Following this path, the structure of this document presents all the different steps fulfilled in order to get to the intended results: –Modification and construction of 3D-printed scale models for the entire UAV, the wing, the tail, and fuselage plus tail models separately. –Conduct wind tunnel experiments for all the separate models (wing and tail alone, and entire aircraft), varying angle of attack and sideslip angle, conducting both static and dynamic experiments1. –Analysis of the results obtained separately and comparison with those obtained with numerical methods. –Identification of sources of error and determination of possible changes in the test methodologies in order to improve the results. 1Static experiments: Experiments conducted for constant angle of attack;Dynamic experiments: Experiments conducted varying angle of attack at a really slow rate such that the experiments can be considered almost stationery. For further information, see section 3. Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 3 The work here presented serves as the basis for future studies, and establish a starting point for more thorough analysis to be conducted in order to complete the aerodynamic flight envelope model of the VTOL-UAV. 2 Model and Experimental Setup Description In this section a brief description of the model and experimental setup is described. First, the modifications introduced to the full scale model both in size and in order to adequate the test model to the wind tunnel facilities are presented, followed by the description of both the wind tunnel and the measuring equipment. 2.1 Wind Tunnel Model Adjustment The VTOL-UAV here studied is characterized by a wing with forward sweep and dihedral, as well as a v-tail, presenting the the following characteristics ([9], [8]): fuselage length of 171.21 cm, wing span of 250.92 cm, tail span of 129.62 cm wing dihedral of 5.27◦and v-tail dihedral of 29.3◦(see Figure 2). Fig. 2 Full scale EMERGENTIA Prototype The airfoils selected for wing and v-tail are the Bell A821201 and a NACA 0012 for the v-tail respectively. Due to the wind tunnel limitations presented in section 2.2, the original model is scaled to a 1:4 ratio. Also in Figure 9 the body fixed and wind reference systems definitions to determine the angle of attack and sideslip angles is introduced. The original full size model had to be scaled in order to conduct the wind tunnel experiments and two criteria were considered to select the appropriate scaling factor: the wind tunnel blockage factor and the size limitations of the 3D-printer that would be use to print all the parts. With these two criteria in mind, a 1:4 scale is chosen in order to make the blocking effects negligible (see section 4 for further details). The final dimensions of the wind tunnel test aircraft are shown in Table 1. Several mod- Fus.length Wingspan Wing surface Dihedral Max. heigth [cm] [cm] [cm2][o][cm] 42.3 62.73 310 5.27 6.31 Table 1 Aircraft dimensions 1:4 scale 4 Victoria M. Prieto, Sergio Esteban ifications were conducted on the original design in order to accommodate the capability of characterize the aerodynamic properties for the wing, v-tail and a complete aircraft separately. These modifications were conducted having in mind that it was desired to minimize the number of elements to be manufactured in order to complete all wind tunnel experiments. For that purpose a series of generic coupling systems were designed in order to enhance the interchangeability between all the different aerodynamic models. Two sets of coupling systems were designed: couplings for the aerodynamic models, and couplings for the measuring devices. The modifications were made using the CATIA V5R19. Figure 4(a) depicts the CAD designs for the aerodynamic model couplings, where, from left to right in the top row depicts the tail and wing coupling structures, while in the bottom shows how the v-tail and wing would be coupled with the coupling structure. Figure 4(b) shows the final 3D-printed products. The material used for the 3D printing was PLA (Polyactic Acid), a thermoplastic aliphatic polyester that, together with ABS (Acrylonitrile Butadiene Styrene), is the most used for 3D printing in what referes to FDM (Fused Deposition Modeling). Although both materials were available for the 3D printing process, ABS is most recommended for high temperature performance and, for that same reason, its use is more complicated in terms of temperature required for its complete melting and also its posterior deposition. Fig. 3 Filling example for 3D printing For that reason, it was decided that PLA was a better choice for these printed models used for these studies, since the work 3D printer temperature was not going to reach high values. Specifically, the melting temperature for the PLA used was 210 ±10Cwith a filament diameter of 1.75mm (from Smart Materials 3D, see [5] and [4]). In what refers to 3D printing specifications, a filling percentage of 20% together with a grid pattern were chosen, values that allow a good mechanical behaviour without waste of material (see figure 3). Also, a printing speed value of 90% of the maximum speed allowed was taken in order to improve the surface quality of the model. Further printing specifications can be found in [19]. On the other side, the couplings for the measuring devices allow to connect the aerodynamic models with the load cell, and is formed by a circular stainless steel section bar with a diameter of 10 mm and a length of 800 mm attached to the transducer by means of a circular plate screwed with four screws (see top left corner of Figure ). A more detailed view of how the aerodynamic models are connected through their measuring devices through the different coupling systems can be seen in Figures 7 and 8. 2.2 Description of the Wind Tunnel and Measuring Setup The wind tunnel is an open circuit one of 23.52 meters length, divided into three sections: contraction chamber, test chamber and diffuser. The test chamber, which is the most important area for the experiments, has a length of 5 meters and a rectangular section 1.4 meters wide and 1.8 meters high. In Figure 5, a panoramic view of the tunnel can be observed, where the contraction chamber (depicted in yellow), the test area (depicted in red) and the diffuser (depicted in green) from left to right, respectively. The maximum design speed of the tunnel is 30 m/s, speed that is achieved thanks to 9 fans (HCT-9- 4T-7.5) positioned equidistant in a 3x3 matrix shape (see Figure 7). Wind tunnel airspeed is given by a potentiometer that regulates the power output of the 9 fans. The true airspeed is measured with a calibrated pitot tube and a pressure indicator Druck DPI 150. Characteristic speeds used for the experiments are depicted in Table 1. A characteristic element of the wind tunnel is the positioning system that allows to vary the angle of attack within the tests. This system consists on two rotating devices each independently moved by an electric motor and controlled by a MicroTest control system (see detail on the walls of the wind tunnel test section in Figure 6(a)), which can be directed from a computer. On the other hand, in order to carry out the measurement of the forces and moments applied to the model, a force and momentum transducer by ATI Industrial Automation [1], called ATI Gamma (see Figure 6(b)) is selected with the following specifications: forces in the x and y axis Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 5 (a) (b) Fig. 4 Coupling modifications of the aerodynamic models Fig. 5 Panoramic view of the wind tunnel range Fx,y±1200 N, forces in the z axis Fz±4100 N, torque in the x and y axis Tx,y±79 Nm, and torque in the z-axis Tz±82 Nm. The data is collected using a custom made interface using LabView software that process the data using a acquisition board which provides the values of forces and moments measured in the transducer in a 6xN matrix written in a .txt file, with the 6 elements of each one of the rows of the matrix describing the load cell measurement of forces and moments in the x, y and z direction, and Nrepresenting the number of lines in the matrix, having the dimensions associated to the frequency used during the acquisition process. (a) (b) Fig. 6 Detail of the wind-tunnel rotating devices and the load cell. 6 Victoria M. Prieto, Sergio Esteban VAVBVCVDVEVF Test values (1:4) [m/s] 5.69 8.47 10.69 13.61 16.75 19.44 Reynolds values 3.48 ·1055.18 ·1056.54 ·1058.32 ·1051.02 ·1061.19 ·106 Table 2 Characteristic speeds and Reynolds values for the wind tunnel The operational range of speed values for the UAV (scale 1:1) for the cruise flight regime are approximated to be around 25-30 m/s, while the expected values during transition flight to hoover range from 0 to 5 m/s. This means that in order to test the model (scale 1:4) in wind tunnel, the achieved speeds in the test section should reach values exceeding 100-120 m/s, which corresponds to a Reynolds between 2.45 ·107and 2.93 ·107. Due to the wind tunnel speed limitations, it was decided to investigate the transition range, throughout a large angle of attack and side-slip envelope which, in addition, we found that there is not too much literature regarding studies of these regimes for this type of aircraft. In fact, these studies are limited to high angles of attack experiments for turbine blades [22,20] and studies on the effect of tail dihedral angle on lateral directional stability due to sideslip angles [18]. Given the wind tunnel physical speed limitations (maximum speed available of 30 m/s), and considering a safety margin in order to avoid possible structure failure, which means not reaching speeds over 20 m/s, the range of test speeds was defined between 5.69 (∼5 km/h) and 19.44 (∼17.5 km/h), as seen in table 2. These velocities allow to conduct a detailed aerodynamic map around the transition flight conditions (including large angles of attack and side-slip) which are of large interest considering the complexity of controlling the aircraft flying during these phases. In addition, the study of these speeds allow to model the influence of the wind gusts affecting the aerodynamic surfaces during hoovering conditions. With the objective of extending the aerodynamic studies to the cruise conditions, further experiments will be conducted in the future using the wind tunnel facilities at the Universidade Federal de Minas Gerais in Brazil which has available a circular wind tunnel capable of reaching speeds above 90 m/s and with a test section that can fit the scale 1:4 model. Although six different velocities have been studied, it is selected as characteristic speed the VC=10.69m/s. This choice is made after observing oscillations for higher speed values at angles over 50othat might disturb the measures. For this reason, despite the fact that all results have been analysed, only results for the characteristic value will be presented. For further details, see [19]. 3 Experiment Procedure Characterizing the entire flight envelope for both the longitudinal an directional modes is an arduous and complex task, that required large amounts of test for variation of angle of attack, sideslip angle, and velocities. In order to optimize the number of test conducted, for future studies, two types of wind tunnel tests were investigated: static and dynamic. The static tests consisted on static measurements for a few seconds for the corresponding model at a given attitude (either at an angle of attack or sideslip), and at the desired speed. Once the flow was stable, data was recorded for several seconds (see section 4.4 for details in the number of data points recorded). On the other side, the dynamic tests consisted in sweeping at different constant angular rates the aerodynamic test models (using the capability of automated rotating devices described in section 2.1) at a given tunnel speed and for a given attitude (either at an angle of attack or sideslip). The original purpose of the dynamic tests was to investigate the possibility of reducing the time associated to conduct static experiments with constant angle of attack or sideslip for future experiments by using this automated methodology. This was considered interesting to investigate since it was estimated that more than 800 test were conducted, since the number of experiments, considered within a matrix of varying cases, grows rapidly: i.e. variations in angle of attack, sideslip angle, wind tunnel speeds, varying sweeping rates, and different aerodynamic models. Table 3 describes the experiments conducted for all the aerodynamic models considering the type of angle analyzed and its range, the velocities studied, the angle lapse discretization for the given range, and the sweep speed. It can be seen that for the angle of attack experiments all the speeds are considered, while for sideslip only the first three ones are tested (VA,VBand VC). As for the angles, the specific ranges for each test are shown in the following table. While in static tests there are difference of lapses between angle of attack and sideslip experiments, dynamic ones are all made with the same sweeping velocity. Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 7 Angle case Angle range Velocity range Lapse Sweep speed α−30o- 90oVA-VD5o10o/min −30o- 40oVE&VF β0o- 50oVA-VC10o10o/min Table 3 Velocity and angle ranges for static and sweep tests The layout of the experiment setup consists on connecting the aerodynamic models and their coupling systems (see section 2.1) to the rotating mechanisms in the walls of the wind tunnel. In order to do so, a structure used in previous projects, [14], has been reused. This structure is made up of two steel bars bolted to the rotating walls of the tunnel and whose central section is a 10 mm high welded metal plate, as seen in Figures 7 and 8. Three couplings have been added to this plate that fix the load cell with the welded metal plate finishing the connection. A detail of two of these attachments can be seen in Figure 9. Thanks to the symmetry of the link between bar and load cell, it is possible to study both the angle of attack and the sideslip by simply rotating 90 degrees the aerodynamic coupling within the bar. Figures 7 and 8 show the detail of the rotation of the final models mounted for the study of variation of angle of attack and for the variation of the slip angle. (a) (b) (c) Fig. 7 Angle of attack layout. Aircraft, wing and tail from left to right (a) (b) (c) Fig. 8 Sideslip layout. Aircraft, wing and tail from left to right 8 Victoria M. Prieto, Sergio Esteban 4 Corrections and Data Processing The purpose of this chapter is to explain in detail all the corrections that have been considered to ensure that the results obtained in the wind tunnel tests are comparable to reality. For this purpose several types of corrections will be taken into account: those associated to wind tunnel corrections, in this case open tunnel with closed section, aerodynamic and weight corrections, and the possible sources of errors associated to the measuring devices and the experiment setup. 4.1 Wind Tunnel Corrections In wind tunnel tests, the flow is limited by its walls, unlike the outdoor flight. This generates interference between the object of study and the walls of the tunnel, a fact that must be taken into account in the design of the models, as previously mentioned. In addition, certain corrections should be considered to compensate these effects. All corrections considered are as follows, whose definitions have been taken from [7]. Horizontal Buoyancy It refers to a variation in static pressure throughout the test section when there is no model on it. In many wind tunnels its value is not zero and produces a drag force analogous to the hydrostatic force on objects in a stationary fluid with a uniform gravitational field. The increase in resistance due to horizontal buoyancy is quantified from Glauert’s expression [13], being λ3the body shape factor for three dimensional bodies and tits maximum thickness. ∆DB=−π 4λ3t3dp dlt (1) dp dlt =−k(ρ/2)V2 B(2) Also, prepresents the static pressure in [kg/m2],Vthe air speed in [m/s],land Bthe wind tunnel length and width, respectively, in [m], and kis a factor that takes a value between 0.016 −0.04. In these wind tunnel tests, two characteristic speeds have been taken: VCand VF. Solid Blockage It refers to the increase in speed produced in the test section by decreasing the air passage section due to the presence of an object. This block depends exclusively on the relationship discussed above: test section - frontal area of the model. The speed increase experienced in this case is determined with different expressions depending on whether it is a wing (3) or a body of revolution (4). Hereby, these are the expressions that allow to calculate these blockage effects in a wing and a fuselage, where, taking the fuselage as a revolution body, the second expression given is considered appropriate, εsb,W=K1τ1 VW C3 2 (3) εsb,B=K3τ1 VB C3 2 (4) where K1and K3are body shape factors obtained from [7] and τ1is a factor depending on the tunnel test section shape and the model span to tunnel width ratio. On the other hand, Crepresents the test chamber section area and VW,VBthe volume of the wing/tail and fuselage. Solid blockage for the entire aircraft is calculated as the summatory of the three parts that compose it: wing, tail and fuselage. Wake Blockage It is a result of the finite size of the body’s wake and has an effect similar to the solid block. However, it is more complicated to characterize since the size of the wake is a function of the body shape and the relationship between the wake area and the tunnel. The larger the size of the wake, the higher the correction factor, which implies an increase in resistance. In closed Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 15 iter ClCdx-velocity y-velocity z-velocity k−ε217 0.12657 0.10555 4.1472 ·10−71.1885 ·10−72.5309 ·10−7 k−ω715 0.15681 0.14499 2.6795 ·10−72.1432 ·10−72.4914 ·10−7 Table 5 Values for standard k−εand k−ωwith SST turbulence models for a representative case of α=10o 6 Results The results obtained after the corrections and the data processing specified before, are presented in this section. Also, these experimental results are compared to the simulations previously carried out in CFD in the case of the complete aircraft, allowing the validation of the model. For simplicity and conciseness of the article, all results presented from now on correspond to static experiments, and only some representative cases are presented. For further information consult [19]. First of all, a comparison of the experimental results for the characteristic speed VCis presented for the three elements studied (plane, wing and tail), as shown in Figures 17 and 18. (a) (b) (c) Fig. 17 Lift, drag and lateral force coefficients for aircraft, wing and tail (a) (b) Fig. 18 Pitch and yaw moment coefficients for aircraft, wing and tail In all cases it is possible to observe that the lift coefficient increases as the angle of attack does (Figure 17(a)). Around 30-40 degrees, there is a small decrease in lift coefficient that anticipates the stall region. Beyond this stall region, the lift coefficient still remains large and slowly decreases its magnitude as the angle of attack continues increasing. 16 Victoria M. Prieto, Sergio Esteban It is observed that lift coefficient values (Figure 17(a)) are higher for the aircraft compared to the wing and tail separately. It is believed that the reason behind this effect might be the shape of the fuselage (side-pods) that results in the generation of vortex that contribute delaying the occurrence of the deep stall region. The fuselage was designed having in mind two objectives: a more functional one such that the side-pods would allow stabilizing the aircraft while in the ground keeping the model from rolling once landed, and eliminating the need of landing gear. A secondary objective was to delay flow detachment at extreme angles of attacks, in special considering the aerodynamic complexity encountered during the transition phase, hence trying to minimize the expected influence by the large fuselage. For that reason it was decided to introduce side-pod alike shapes on each side of the fuselage to help creating a similar effect to the one created by the LEX (Leading-edge extensions) that some aircraft use in high angle of attack configurations with the objective to improve the airflow at high angles of attack and low air-speeds, [21,6,15]. Although no expectations of significant contribution of the side-pod to the lift was expected, it can be seen in Figure 17(a) that the fuselage introduces a significant contribution to the total lift for each alpha value (not only after stall). These results are also consistent with the comparison studies of numerical and experimental results for aircraft (CL,CD,CY) presented in Figure 21. Despite this initial believe, this effect needs to be further investigated with flow visualization techniques that will be introduced in the future fuselage wind tunnel tests to clarify its contribution. When analyzing the drag of the entire aircraft, it can be seen that, as expected, it is greater than the other two parts (wing and v-tail alone), and decreases when the angle gets closer to zero as the effective area is reduced (see Figure 17(b)). Looking in detail to the drag contribution of the fuselage 17(b), it can be also seen that, although the drag increases, it does so in a quite linearly manner in the alpha range between 20◦−70◦, despite a quadratic evolution would be expected at large angles of attack due to the large flow detachments, since, as depicted by the lift behaviour described above, a weaker drag increase would be expected to be consistent with the explanation increase in lift contribution by the fuselage due to the side-pods helping to control the unstable flow turbulence, hence reducing the drag growth. These results are also consistent with the comparison studies of numerical and experimental results for aircraft (CL,CD,CY) presented in Figure 21. As mentioned earlier, this phenomenon is worthy of investigating with experimental flow visualization techniques that will be conducted in future studies. As for the pitching moment coefficient, in all cases it has a negative slope, then it is stable. The moment generated on the aircraft is greater than in the cases of wing and tail because the contributions of both are added. In these cases, there is a negative moment, that is to say, for positive angles of attack and vice versa, a nose-up pitching moment for negative angles of attack (see Figure 18(a)). In the case of lateral force (see Figure 17(c)), the contributions of wing and tail together with fuselage give rise to a positive force (in the direction of the rotation of the model) ascending, as it happens with the tail but not with the wing, which decreases and takes negative values (in the opposite direction to the turn). Finally, the yaw moment (Figure 18(b)), with a tendency opposite to the lateral force, gives rise to a recovery moment in all the studied cases. Some representative experiments are presented in Figures 19 and 20. (a) (b) (c) Fig. 19 Experimental results for various velocities (Force coefficients) Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 17 (a) Fig. 20 Experimental results for various velocities (Moment coefficients) Finally, some numerical results are presented in Figure 21. It is observed that both lift and drag take values very similar to the real case measured in the wind tunnel. There are small discrepancies due, with great probability, to the ideal conditions of symmetry and constant flow. Likewise, within the tunnel there was a margin of ±1.5oin the load cell that can contribute to the differences. For the rest, it is demonstrated that the real behavior before variations of angle of attack is very similar to the one found numerically, so the model is validated. Also, some views of the velocity flow around the aircraft are presented (Figure 22). (a) (b) (c) Fig. 21 Numerical and experimental results comparison for aircraft (CL,CD,CY) 7 Conclusions and Future Recommendations The article presents wind tunnel experiments conducted on a convertible VTOL-UAV in order to characterize the entire full aerodynamic flight envelope through the extended analysis of angle of attack and sideslip angles. The obtained results provide an initial estimate of the aerodynamic maps for the wing, tail and complete aircraft models that will serve as bench mark to develop the high fidelity simulator models, hence accomplishing the initial objectives. 18 Victoria M. Prieto, Sergio Esteban (a) (b) (c) Fig. 22 Velocity flow visualization for CFD analysis. (α=5o,α=90o,β=30ofrom left to right) The results present a good performance of the entire aircraft both for the lift, and drag coefficient after the stall region which can be attributed to the side-pod LEX alike shapes that seem to help delaying the occurrence of deep stall phenomena, which is worthy to investigate in more detail with experimental flow visualization techniques that will be conducted in future studies. In addition several lessons can be extracted from the presented results. In spite of all the problems encountered during the realization of the sweep tests that invalid the results obtained (see section 4.4), these experiments seem promising for future studies as they would mean a considerable reduction of number of experiments and time. For this purpose, structural analysis are being conducted on the structure and the coupling attachments to identify the source of damage and, if necessary, redesign in order to be able to continue studying the feasibility of the sweep test. For that matter, and in order to eliminate the possibility that the failure occurred because of an unstable dynamic behaviour due to frequencies introduced during the sweep test in the system that might have been close to the natural frequencies of the setup, a throughout modal analysis is also going to be conducted for a large number of rotational speeds. Despite the problems that have arisen and the differences mentioned, the behavior of the aircraft as a whole and of each of the aerodynamic models (wing and v-tail) in the static tests seems to be adequate and resemble the CFD results, so the main objective of this project has been fulfilled. Future project lines derived from this work are shown below: –Analysis of the variation of roll angle and sideslip-attack angle coupling. –Structural study of the load cell and improvements in the wind tunnel. –Repetition of these experiments with a more exhaustive study (more angles and speed values) to validate the model. 8 Acknowledgments This research has been supported by the Ministerio de Ciencia, Innovaci´ on y Universidades grant RTI2018-101519-A-I00. 9 Conflict of Interest Statement On behalf of all authors, the corresponding author states that there is no conflict of interest. References 1. F/t sensor gamma,ati industrial automation, in-line documentation,. Documentation. URL: https://www.ati-ia.com/products/ft/ft_ models.aspx?id=Gamma [cited September 2020] 2. Law of the wall. Image. URL: http://brennen.caltech.edu/fluidbook/basicfluiddynamics/turbulence/lawofthewall.pdf [cited 19 October 2020] 3. The numerical computation of turbulent flows. Computer Methods in Applied Mechanics and Engineering 3 4. Sculpteo. pla definition. Web. URL: https://www.sculpteo.com/en/glossary/pla-definition/ [cited 20 October 2020] 5. Smart materials 3d. Web. URL: https://www.smartmaterials3d.com/pla-filamento [cited 20 October 2020] Wind Tunnel Aerodynamic Study of a VTOL UAV For Wide Range of Angles of Attack and Sideslip Angles 19 6. Leading-edge vortex formation and transient lift generation on a revolving wing at low reynolds number. Aerospace Science and Technology 97, 105589 (2020). DOI 10.1016/j.ast.2019.105589 7. Barlow J.B., P.A., Rae, W.: Low-speed wind tunnel testing. Wiley-Interscience (1999) 8. Esteban S., e.a.: Desarrollo de una aeronave convertible no-tripulada para despliegue r´ apido y eficiente en situaciones de emergencia: Proyecto emergentia. In: IV Congreso sobre las Aplicaciones de los DRONES a la Ingenier´ ıa Civil – CivilDRON’18 (2018) 9. Esteban S., e.a.: Development of unmanned convertible aircraft for rapid and efficient deployment in emergency situations: Project emergentia. In: DRONE Berlin 2017. DEMACH Event- und Veranstaltungs GmbH (2018) 10. Frank I., J.J.H.D.: Wind-tunnel static and free-flight investigation of high-angle-of-attack stability and control characteristics of a model of the EA-6B airplane. Nasa technical paper 3194, Hampton, Virginia (1992) 11. Gatto, A.: Application of a pendulum support test rig for aircraft stability derivative estimation. Journal of Aircraft 46(3), 927–934 (2009). DOI 10.2514/1.38916 12. Gatto, A., Lowenberg, M.H.: Evaluation of a three degree of freedom test rig for stability derivative estimation. Journal of Aircraft 43(6), 1747–1761 (2006). DOI 10.2514/1.19821. URL https://doi.org/10.2514/1.19821 13. Glauert, H.: Wind Tunnel Interference on Wings, Bodies and Airscrews. HM. STATIONERY OFFICE (1933) 14. G´ omez Fuster, I.: Plataforma para la medici´ on de fuerzas y momentos aerodin´ amicos de modelos a escala en un t´ unel de viento. Proyecto fin de carrera, Sevilla (2013). URL: http://aero.us.es/sesteban/pdf/PFC/Windtunnel/PFC_Isabel/PFC_IGF.pdf [cited 27 October 2020] 15. Mai, H., Dietz, G., Geißler, W., Richter, K., Bosbach, J., Richard, H., de Groot, K.: Dynamic stall control by leading edge vortex generators. Journal of the American Helicopter Society 53, 26–36 (2008). DOI 10.4050/JAHS.53.26 16. Menter, F., Kuntz, M., Langtry, R.: Ten years of industrial experience with the sst turbulence model. Heat and Mass Transfer 4(2003) 17. Menter, F.R.: Two-equation eddy-viscosity turbulence models for engineering applications. AIAA Journal 32(8), 1598–1605 (1994). DOI 10.2514/3.12149 18. Nur Amalina Musa Shuhaimi Mansor, A.A., Man, M.H.C.: Effect of tail dihedral angle on lateral directional stability due to sideslip angles. In: 53rd AIAA Aerospace Sciences Meeting, Kissimmee, Florida. AIAA (2015). DOI 10.2514/6.2015-1498 19. Prieto Rueda, V.: Estudio aerodin´ amico en t´ unel de viento de un uav vtol-cp para misiones sar. Trabajao fin de grado, Sevilla (2019). URL: http://aero.us.es/sesteban/pdf/PFC/VTOL/Aerodinamica/TFG_Victoria_Prieto/TFG_Victoria_Prieto.pdf [cited 27 October 2020] 20. Seak Hoo Lee, Y.O.H.: Experimental Investigation of High-Angle-of-Attack Aerodynamics of Low-Aspect-Ratio Rectangular Wings Configured with NACA0012 Airfoil Section. International Journal of Aeronautical and Space Sciences (21), 303–314 (2019) 21. Sutrisno, S., Deendarlianto, D., Rohmat, T.A., Wibowo, S.B., Iswahyudi, S.: Vortex Dynamics Analysis of Straight-Body-Type-Fuselage Fighter Using CFD Simulation. Journal of Aerospace Technology and Management 22. Timmer, W., Rooij, R.: Some aspects of high angle-of-attack flow on airfoils for wind turbine application (2001) 23. Wilcox, D.: Turbulence Modeling for CFD. No. v. 1 in Turbulence Modeling for CFD