key: cord-0058852-3pjxhwd5 authors: Natale, Nunzio; Salomone, Teresa; De Stefano, Giuliano; Piccolo, Antonio title: CFD Prediction of Aircraft Control Surfaces Aerodynamics date: 2020-08-24 journal: Computational Science and Its Applications - ICCSA 2020 DOI: 10.1007/978-3-030-58799-4_6 sha: ec2f52a3a31c4d2ebbb7c0f997aadbd2661b1254 doc_id: 58852 cord_uid: 3pjxhwd5 The effectiveness of prototypical control surfaces for a modern regional transport commercial aircraft is examined by means of numerical simulations. The virtual experiments are performed in operational conditions by resolving the mean turbulent flow field around a suitable model of the whole aircraft. The Reynolds-averaged Navier-Stokes modelling approach is used, where the governing equations are solved with a finite volume-based numerical technique. The aerodynamic performance of the flight control surfaces, during an hypothetical conceptual design phase, is evaluated by conducting simulations at different deflections. The present computational modelling approach is verified to have good practical potential by making a comparison with reference industrial data. Recent progresses in meshing technology, convergence acceleration and calculation performance have allowed Computational Fluid Dynamics (CFD) methods to be intensively used to generate relevant aerodynamic databases, which are particularly useful in the pre-design process of new aircrafts [1, 2] . Present trends in industrial aerodynamic design towards cost reduction for product development call for an accurate prediction of the control surfaces aerodynamics, in order to properly arrange the flight control system. The estimation of handling qualities and hinge moments induced by the deflection of the different control surfaces plays a key role in the aircraft sizing process, with a strong impact on the aircraft weight prediction. However, for flows around deployed control surfaces, the complex aerodynamics, as well as the importance of the flight envelope to be covered, makes the numerical simulations particularly challenging. Differently from typical CFD studies that are performed by employing simplified generic wing-body configurations [3, 4] , in this study, the computational evaluation of the prototypical configuration of a real regional transport aircraft is performed, with a detailed analysis of the local aerodynamics of the flight control surfaces. The simulations allow to determine the variation of the aerodynamic loads, depending on the deflections of the control surfaces, and thus examine the effectiveness of the control process. Specifically, the effect of ailerons and elevators is studied at a typical flight condition, by means of a number of different calculations. The detailed air flow field is simulated by employing a fully three-dimensional CFD method, while following the Reynolds-averaged Navier-Stokes (RANS) approach. The numerical simulations are conducted using the solver ANSYS Fluent, which is commonly and successfully employed for building virtual wind tunnels in industrial aerodynamics research [5] [6] [7] [8] . Indeed, this approach is followed in similar works investigating the aerodynamics of innovative regional aircraft models [9, 10] . The feasibility of the proposed computational model for the simulation of complex realistic configurations is assessed through dedicated comparisons with analytical/experimental data provided by the Leonardo Aircraft industrial researchers. The preliminary prediction of the global aerodynamic coefficients of interest is obtained using lower fidelity methods, by using either the Vortex Lattice Method (VLM) or a semi-empirical approach [11] . VLM solvers are ideally suited for the preliminary aircraft design environment, where they can be used to predict loads, stability and control data. Some aerodynamic characteristics as lift curve slope, induced drag and lift distribution are obtained through modelling the wing by means of horseshoe vortices distributed along both span and chord, while neglecting the effects of thickness and air viscosity [12] . Furthermore, the aerodynamics of the control surfaces is more accurately predicted by solving the RANS governing equations, which describe the mean turbulent flow field around the aircraft. In this work, the Spalart-Allmaras oneequation model is used, where the turbulence closure is achieved by solving an additional evolution equation for a modified eddy viscosity variable [13] . The turbulent boundary layer is modeled using a wall-function approach. This way, the wall boundary conditions are implemented so that relatively coarser meshes can be used in the calculations. The equations governing the RANS models, which are not reported here for brevity, can be found, for instance, in [14] . The RANS method is commonly used for the simulation of both compressible and incompressible aerodynamic flows, e.g. [15, 16] . Complete Aircraft. The geometric model for the regional transport aircraft under investigation is comprised of fuselage, wing, empennage, and nacelle groups. Some important geometrical data of the aircraft model are given in Table 1 . Here, the different geometrical parameters are expressed in nondimensional form, by using the wing area A and the wing span b for normalization. The taper ratio λ is defined as the ratio between the tip chord and the root chord of the wing. Note that the actual reference values are not reported in the table, because they are not classified for publication. A baseline simulation for the clean configuration is first performed. Practically, the gaps between the main aircraft components and the stowed control surfaces are not considered, while the projections of the surfaces edges are maintained, so that the different contributions of the different elements can be evaluated. A hybrid mesh with inflation layers is employed for all the CFD calculations. The numerical simulations for the clean configuration are carried out only for symmetric flow conditions and, thus, half the geometric model is used, while imposing proper symmetry boundary conditions. The FV mesh that is actually used corresponds to the finest mesh produced through a proper sensitivity analysis, being made up of about 7 million elements. Moreover, due to the wall function approach, a suitable first layer thickness is selected, which allows to properly model the boundary layer flow region. The mesh quality, as it holds for all the FV meshes used in this work, can be considered pretty good, especially considering the complexity of the geometrical model. In fact, the maximum skewness of the mesh is less than 0.9 for all the cases, which is within the prescribed limits. The same occurs for other mesh quality parameters, such as the aspect ratio and the orthogonal quality. The frontal view of the aircraft geometry for the clean configuration is presented in Fig. 1 , along with the associated surface mesh. Note that the orientation of the reference body axes is such that the x-axis points backward, the y-axis points to the right wing, and the z-axis points upward. The CAD models for the control surfaces are developed by exploiting the geometrical data provided by industrial aerodynamicists. The analysis is conducted by deflecting only one of the control surfaces, while maintaining the others with no deflection. An important aspect to be considered is regarding the gaps between the moving surfaces and the correspondent main components. For each control surface, the upper and lower gaps are closed, while the lateral gaps are kept. Some geometrical data associated to the aircraft control surfaces system are summarized in Table 2 , and illustrated in Fig. 2 . The characteristic high wing T-tail empennage configuration of the regional transport aircraft under study is illustrated in Fig. 3 , where z h −z w and l h represent, respectively, the vertical and horizontal distances between the horizontal tailplane and the aircraft wing. The incidence of the horizontal tailplane is represented by i h , which is the negative setting angle of the tailplane chord c h with respect to the x-axis direction, as depicted in Fig. 4 . Ailerons. The CAD model that is developed to study the aerodynamic effect of the ailerons is presented in Figure 5 , along with the associated surface mesh. The mesh features are corresponding to those ones considered for the clean configuration, while imposing two opposite deflection angles for the two ailerons, which are δ a = ±10 • . Since a symmetrical model can not be employed in this case, the mesh generation is simplified by separately considering the two specular computational sub-domains, while putting a suitable interface between them in Elevators. The model for studying the aerodynamic effect of the elevators is fully consistent with the approach used above. The geometrical characteristics of the elevators are presented in Table 2 . In this study, a symmetrical mesh can be used due to the fact that no simulation is performed with non zero sideslip angle. The CAD model for the right elevator is illustrated in Fig. 7 , along with the associated surface mesh, for a given positive deflection angle that is δ e = 10 • . The Isolated Wing. The loading distribution along the spanwise direction is evaluated through the strip method, by developing a suitable custom field function. Namely, the wing is divided into a series of strips on which pressure and viscous based loads are computed in terms of the product cC l , where c and C l stand for the local chord length and the section lift coefficient. As initial verification, the spanwise loading distribution obtained by an inviscid CFD calculation is compared to the VLM-based solution, for a flight condition corresponding to a lift coefficient C L = 0.48. As demonstrated in Fig. 8 , where the chord length is normalized by the mean aerodynamic chord (c ref = c), the agreement between the two different solutions is fully acceptable. In this figure, as well as in the following, the abscissa Y = y/(b/2) represents the normalized coordinate along the spanwise direction. Complete Aircraft. The model of the complete aircraft in clean configuration is used for a number of numerical tests, whose results are exploited as reference for the following experiments with aerodynamic control surfaces deflected. In the following discussion, the main aerodynamic coefficients and the spanwise loading distributions are examined. In particular, the three different moment coefficients are defined as C r (rolling), C m (pitching), and C n (yawing), where it is assumed C r > 0 when the rolling tends to put the right wing down, C m > 0 when the pitching tends to put the aircraft nose up, and C n > 0 when the yawing tends to put the aircraft nose right. Note that, given the body axes reference introduced above, it practically holds C r = −C Mx , C m = C My , and C n = −C Mz , where M x , M y and M z are the moment components along the thee different body-axis directions. The reference point for the moment evaluation is set coincident with the point at one quarter of the mean aerodynamic chord. A number of steady simulations are conducted at a given flight condition, which is expressed in terms of Mach number and Reynolds number that is Re ∞ = 11.8 × 10 6 . The aerodynamic loadings on the aircraft are obtained by integrating the predicted surfaces forces. The drag polar for the clean configuration is illustrated in Fig. 9 , compared with the theoretical parabolic polar, that is where C D0 stands for the minimum drag coefficient, AR is the aspect ratio and e is the Oswald factor. The present CFD calculations provide C D0 equal to 340 drag counts, corresponding to a lift coefficient C L of about 0.2. The parabolic curve fitting the numerical data leads to estimate e = 0.9 for the Oswald factor. Apparently, the lift coefficient is well predicted for a reasonable operative range of the AoA parameter. The pitching moment coefficient is reported on the left side of Fig. 10 , where the negative slope obeys the longitudinal stability criteria. The neutral point has been predicted at 53% of the mean aerodynamic chord, which is consistent with experimental data [17] . The increased C m slope at high AoA is due to the pendular stability, as is typical for high wing aircrafts [18] . The breakdown of the various contributions to the pitching moment is illustrated on the right side of Fig. 10 . The addition of the vertical tail to the wing-body (WB) combination does not affect C L and C m , whereas longitudinal stability is introduced by the horizontal tailplane. The body-tailplanes configuration (BVH) is stable but does not produce useful lift. The complete aircraft configuration has the highest C L slope due to the additional lift provided by the horizontal tailplane, yet it represents an untrimmed configuration. In the following, in order to show the capability of the proposed computational modelling, the aerodynamic effect of ailerons and elevators deflection is examined. The results are presented and discussed for some suitable deflection angles. Ailerons. The aerodynamic effect of the ailerons is presented for deflection angles equal to δ a = ±10 • , with the right aileron deflected downward. The predicted spanwise loading distribution is presented in Fig. 11 , where the loading is calculated up to the sections immediately before the winglet (Y = ±1). When making a comparison with the result of the baseline simulation with no deflection, the influence of the ailerons is apparent. In Fig. 12 , the rolling moment coefficient variation with the lift coefficient is reported, compared to VLM results. A loss in ailerons effectiveness for C L > 0.9 is observed in the CFD solution, while the VLM method is not able to predict it. The corresponding drag breakdown is illustrated in Fig. 13 , where each contribution is expressed in percentage to the total value. Finally, the adverse yaw effect is quantified by the derivative C nδa , which is equal to 2.6 × 10 −3 rad −1 at zero AoA and rises up to 1.6 × 10 −2 rad −1 for an AoA of 7.5 • . Elevators. The aerodynamic effect of the elevators is presented for two different deflection angles that are δ e = ±10 • . Figure 14 shows the spanwise loading distribution for the wing and the horizontal tailplane, for both deflections. The presence of the vertical tailplane moves the maximum loading slightly away from the symmetry plane. In particular, the interaction seems to be stronger on the lower surface of the horizontal tailplane. In Fig. 15 , the spanwise loading distribution on the horizontal tailplane with elevator deflected, compared with the baseline solution with no deflection, is reported. The corresponding pitching moment coefficient plots are presented on the left side of Fig. 16 . Apparently, the horizontal tailplane provides a negative lift when the elevator is not deflected, due to its initial negative incidence with respect to the aircraft centerline (i h < 0) and the downwash effect from the wing. Deflecting the elevator by −10 • increases the absolute value of the negative lift. On the other hand, deflecting the elevator by +10 • provides a positive lift, that is in absolute value lower than the previous case. The drag polar obtained for δ e = −10 • is shown on the right side of Fig. 16 . It is worth noting that the negative deflection of the elevator leads to a relevant increment of 78 drag counts for C D0 . By looking at the drag breakdown, which is reported in Fig. 17 , it is interesting to note that deflecting the elevator by −10 • (up) leads to higher drag and influences the vertical tailplane as well. This is due to the fact that the elevator leads to higher (negative) load in this case. The pitching moment coefficient variation for the deflection angle δ e = +10 • is equal to ΔC m = −0.536, which is comparable to the value obtained using a semi-empirical method that is −0.61 [11] . When making the same analysis for the lift coefficient increment, at zero AoA, it holds ΔC L = 0.109, while the semi-empirical method providing a value of 0.1. For all the aerodynamic derivatives calculated in the present study, a remarkable matching was found with respect to the values generated by Leonardo Aircraft aerodynamicists. These data were provided as restricted classified data and are not available for public diffusion. Finally, as far as the comparison with the VLM solution is concerned, the discrepancy of the results is large, likely due to differences in the geometries and/or meshes employed. This was also found in similar studies [19] . The present study has to be intended as the proof-of-concept, namely, the preliminary development of a CFD based prediction tool in the aerodynamic design of the control surfaces system of a regional transport aircraft. Differently from similar studies, where the CFD results were obtained for generic geometries, this work deals with simplified models directly derived from real geometries provided by an aeronautical manufacturer that is Leonardo Aircraft Company. Several calculations have been performed with the aim of demonstrating the practical potential of the proposed methodology for the prediction of the control surfaces aerodynamics. The analysis is conducted by performing calculations at a given flight condition and for different deflection angles of ailerons and elevators, while examining the modified aerodynamic force and moment coefficients. It is demonstrated that a complete aerodynamic CFD database can be generated, given the prototypical geometry of the control surfaces, in the pre-design process of new aircrafts. The acceptable agreement with the empirical data made available by the industrial partner has been achieved. There remains the possibility of developing more sophisticated computational models for particular industrial applications, depending on the level of accuracy that is required and the available computational resources. Thirty years of development and application of CFD at Boeing Commercial Airplanes On the role and challenges of CFD in the aerospace industry Numerical simulations around wing control surfaces 3D viscous flow analysis on wingbody-aileron-spoiler configurations Detached-eddy simulation of a wide-body commercial aircraft in high-lift configuration Computational evaluation of wind loads on suntracking ground-mounted photovoltaic panel arrays A supersonic jet target for the cross section measurement of the 12C(α, γ)16O reaction with the recoil mass separator ERNA Multidisciplinary design optimization of aircraft wing using commercial software integration Aerodynamic and aeroacoustic investigation of an innovative regional turboprop scaled model: numerical simulations and experiments Computational evaluation of aerodynamic loading on retractable landing-gears Airplane design part VI: preliminary calculation of aerodynamic, thrust and power characteristics Historical evolution of Vortex-Lattice Methods A one-equation turbulence model for aerodynamic flows Turbulence Modeling for CFD Wavelet-based adaptive unsteady Reynolds-averaged Navier-Stokes simulations of wall-bounded compressible turbulent flows Wavelet-based adaptive unsteady Reynolds-averaged turbulence modelling of external flows Wind tunnel testing of a generic regional turboprop aircraft modular model and development of improved design guidelines Airplane Performance Stability and Control Validation of numerical prediction of dynamic derivatives: the DLR-F12 and the Transcruiser test cases Acknowledgements. This work was supported by the Italian Regione Campania under the research project SCAVIR (POR Campania FESR 2014/2020), presented by the Campania Aerospace Technological District (DAC). This support is gratefully acknowledged. The authors would like to thank the Leonardo Aircraft Company for providing the CAD models and the design reference data.