M.Ing. / The aim of this dissertation is to apply numerical aerodynamic principles to the characterization of an alternative stepped aerofoil concept. The accurate and efficient determination of the aerodynamic forces caused by the relative fluid motion and the consequent lift and drag coefficients are essential for the characterization of new aerofoils. The numerical method used is in the form of a Computational Fluid Dynamics code, which integrates the Navier-Stokes equations through finite-volume dictretization principals. A two-dimensional approximate analysis procedure is used together with a two-equation turbulence approximation in the form of the "standard" k-c turbulence model. Available software is used and adapted where applicable. A suitable method for comparing wing section characteristics as a function of profile geometry and attitude is developed in this thesis. This is achieved by first refining a numerical test case and quantifying the influences of model parameters such as grid design, boundary conditions and solution variables. Alternative geometrical aerofoil concepts can then be characterized by employing the same principles. This thesis contains selected results of hundreds such numerical simulations, all of which were necessary to refine the test case and eventually characterize the aerofoils. The proposed wing section geometry, incorporating a rearward-facing step shows some improvement in aerodynamic performance over a standard reference case. Geometrical variations of the step concept are also investigated and can later be used in an optimization procedure. A transient simulation approach is employed for unsteady cases and flow visualization is done in order to learn more about the unique aerodynamic action of the proposed concept. Experimental results obtained in a wind tunnel for the pressure around the investigated aerofoils are used to verify numerical results. Further development in the numerical approach may include the use of additional, more advanced turbulence models. This may allow the research of more complex phenomena such as stall and also broader ranges of Reynolds numbers in more detail. To complete the characterization process, the moment coefficients should also be included.

Ph. D.

Time-averaged properties of a wing-body-junction flow surrounding a cylindrical wing with a 1.5:1 elliptical nose and a NACA 0020 tail have been compared to those for the above wing with a 1.5 inch radius fillet normal to the wing's surface. An attempt was made to determine the effectiveness of the fillet in improving the uniformity of the wing-body junction flow downstream of the wing, and in attenuating the junction vortex. Measurements included oil-flow visualizations and surface-static-pressure measurements of the flattest floor surrounding the wings, and hot-wire anemometer measurements made in the flow downstream of the wing. Calculations of the drag and the volumetric entrainment of free-stream fluid due to the presence of the baseline wing and wing with fillet were performed. The results of these calculations are important criteria used to determine the effectiveness of the fillet as a flow control device. Results show that the vortex is present in each case, and its size is slightly larger for the wing with fillet as compared to the baseline wing. For each test case, the drag and volumetric entrainment of free-stream fluid were the nearly same for the wing with fillet as compared to the baseline wing. It was also found that increases in the boundary-layer thickness cause only small increases in the size of the junction vortex. The 1.5 inch radius fillet does not appear to be a viable flow control device. / Master of Science

The introduction of composite materials and the ability to tailor these materials to improve aerodynamic and structural performance is having a distinct effect upon aircraft design. In order to optimize the efficiency of the design procedure, a design process which is more integrated than the traditional approach is required. Currently the utilization of such design procedures produces enormous computational costs. An ongoing effort to reduce these costs is the development of efficient methods for cross-disciplinary sensitivities and approximate optimization techniques. The present research concentrates on investigating the integrated design optimization of a subsonic, forward-swept transport wing. A modular sensitivity approach for calculating the cross-sensitivity derivatives is employed. These derivatives are then used to guide the optimization process. The optimization process employed is an approximate technique due to the complexity of the analysis procedures. These optimization results are presented and the impact of the modular technique is discussed. / Master of Science

Ph. D.

Ph. D.

A technique to obtain the sensitivity of the static aeroelastic response of a three dimensional wing model is designed and implemented. The formulation is quite general and accepts any aerodynamic and structural analysis capability. A program to combine the discipline level, or local, sensitivities into global sensitivity derivatives is developed. A variety of representations of the wing pressure field are developed and tested to determine the most accurate and efficient scheme for representing the field outside of the aerodynamic code. Chebyshev polynomials are used to globally fit the pressure field. This approach had some difficulties in representing local variations in the field, so a variety of local interpolation polynomial pressure representations are also implemented. These panel based representations use a constant pressure value~ a bilinearly interpolated value, or a biquadratic ally interpolated value. The interpolation polynomial approaches do an excellent job of reducing the numerical problems of the global approach for comparable computational effort. Regardless of the pressure representation used, sensitivity and response results with excellent accuracy have been produced for large integrated quantities such as wing tip deflection and trim angle of attack. The sensitivities of such things as individual generalized displacements have been found with fair accuracy. In general, accuracy is found to be proportional to the relative size of the derivatives to the quantity itself. / Ph. D.

An analysis was presented for predicting lift and drag characteristics of flat-top wing-body combinations at supersonic speeds. These combinations consist of a wing mounted above an expanding body with their apexes being coincident. The assumptions with which the analysis was made are the following: 1. The linear theory was applicable. 2. The leading edge or the wing was coincident or ahead of the body shock. 3. Condition of zero base drag (static pressure at the base equal to the stream static pressure). The analysis was carried out by considering the individual terms which appear in the lift-drag relations separately, and utilizing the most recent theoretical methods to determine them. The analysis was applied to two flat-top wing-body combinations; namely, a semiconical body with an arrow planform wing, and a 3/4 power semibody with a diamond planform wing. For these combinations a free-stream Mach number of 3.35 satisfied the condition for the wing leading edge and the body bow shock to be coincident. To obtain a check on the analysis, the results were compared with experimental data at a Mach number of 3.35. / Master of Science

An experimental and computational investigation of a low-canard high-wing aircraft design has been conducted. The aircraft studied has a canard and wing of similar chord and airfoil section. The canard is approximately half the span of the main wing and both surfaces are untwisted and unswept. Canard incidence with respect to the zero angles of attack line is 4° and the main wing has an incidence of 1° and a dihedral of 3°. Force and moment data were obtained in two separate wind tunnel test entries in the VPI Stability Tunnel. The first of these entries were concerned with longitudinal characteristics while the second dealt primarily with lateral/directional characteristics. Flow visualization was also done in both sets of tests. Lift characteristics showed an apparent onset of stall and then a second rise in the lift curve. The aircraft displayed stable characteristics in both the longitudinal and lateral/directional cases. However, the pitch break at the onset of stall was unstable. The “double peak” lift curve as well as the unstable pitch break have been attributed to the canard tip-vortex interaction with the main wing. Test Reynolds numbers were 260,000 for the first set and 300,000 for the second series of tests. Computational cases were run using both an uncambered vortex lattice method and a general three-dimensional constant doublet and source panel method. Lift curve slope and static margin were obtained from the vortex lattice code and agree well with the experiment. All aerodynamic forces and moments were predicted by the doublet panel method PMARC. Longitudinal data was obtained using a symmetric 3200 panel model while lateral/directional data was taken using a 1600 panel model. Both the lift curve slopes and the pitching moment slopes compare well between the computational cases and the experimental data. The actual values for a given angle of attack, however, differ and remain unexplained. This is possibly due to either canard wing interaction effects, wind-tunnel-model manufacturing flaws, model mount or tunnel installation interference or a data reduction error. / Master of Science

Past research at this university has proven the feasibility of the multi-disciplinary design of a complex system involving the complete interaction of aerodynamics and structural mechanics. Critical to this design process, is the ability to accurately and efficiently calculate the sensitivities of the involved quantities (such as drag and dynamic pressure) with respect to the design variables. These calculations had been addressed in past research, but it was felt that insufficient accuracy had been obtained. The focus of this research was to improve the accuracy of these sensitivity calculations with a thorough investigation of the computational aspects of the problem. These studies led to a more complete understanding of the source of the errors that plagued previous results and provided substantially improved sensitivity calculations. Additional research led to an improvement in the aerodynamic-structural interface which aided in the accuracy of the sensitivity computations. Furthermore, this new interface removed discontinuities in the calculation of the drag which the previous model tended to yield. These improvements were made possible with the application of shape functions in surface deflection analysis, instead of the previous ‘zonal’ approach. Other factors which led to accuracy improvements were changes to the aerodynamic model and the paneling scheme. Final studies with the optimization process demonstrated the ability of the improved sensitivities to accurately approximate the design problem and provided useful results. Additional studies on the optimization process itself provided information on move limit restrictions and various constraint problems. / Master of Science