The dissertation is focused on the large computational costs of integrated multidisciplinary design. Efficient techniques are developed to reduce the computational costs associated with integrated structural-aerodynamic design. First efficient methods for the calculations of the derivatives of the flexibility matrix and the aerodynamic influence coefficient matrix are developed. An adjoint method is used for the flexibility sensitivity, and a perturbation method is used for the aerodynamic sensitivity. Second a sequential optimization algorithm that employs approximate analysis methods is implemented. Finally, a modular sensitivity analysis, corresponding to the abstraction of a system as an assembly of interacting black boxes, is applied. This method was developed for calculating system sensitivity without modifying disciplinary black-box software packages. The modular approach permits the calculation of aeroelastic sensitivities without the expensive calculation of the derivatives of the flexibility matrix and the aerodynamic influence coefficient matrix. / Ph. D.

Parawing vehicles may have unusual values of many of the mass and aerodynamic factors affecting dynamic lateral stability and control. These unusual characteristics are due in large part to the fact that the center of gravity of parawing vehicles is located far below the parawing, whereas conventional aircraft usually have the vertical center-of-gravity location near the plane of the wings. The present thesis is an analytical investigation of the dynamic lateral stability and control of a typical parawing vehicle. The analysis was made using three-degree-of-freedom, rigid body equations of motion. Stability derivatives used in the calculations were obtained from static and dynamic force tests of a parawing model with rigid leading-edge and keel members. The analysis is treated mainly in terms of the effects of vertical center-of-gravity position, since this was found to be the most significant factor affecting the lateral stability and control of the hypothetical vehicle. / Master of Science

A general nonlinear, nonplanar unsteady vortex panel method for potential-flow is developed. The surface is modeled as a collection of triangular elements on which the vorticity vector is piecewise linearly varying. The wake emanates from the sides and trailing edges of the thin lilting surfaces and is modeled as a progressively formed collection of vortex filaments. This model provides a continuous pressure distribution on the surface while allowing the wake to roll up as tightly as needed. The wake position is determined as part of the solution and no prior knowledge of the position or strength is assumed. An adaptive grid technique is used to redistribute the circulation of the vortex filaments of the wake as the wake sheet spreads. The aerodynamic model is coupled with dynamic equations of motion. Forced oscillation tests are conducted on flat rectangular and delta wings. Dynamic tests are performed to predict wing rock of a slender delta wing restricted to one degree of freedom in roll. The aerodynamic/dynamic model is coupled with control laws that govern the motion of flaperons so that a prescribed pitch motion is executed and wing rock is suppressed. ( 300 pages, 107 figures ) / Ph. D.

In many situations the motion of the fluid and the motion of the body must be determined simultaneously and interactively. One example is the phenomenon of subsonic wing rock. A method has been developed that accurately simulates the pitching and rolling motions and accompanying unsteady flowfield for a slender delta wing. The method uses a predictor-corrector technique in conjunction with the general unsteady vortex-lattice method to compute simultaneously the motion of the wing and the flowfield, fully accounting for the dynamic/aerodynamic interaction. For a single degree of freedom in roll, the method predicts the angle of attack at which the symmetric configuration of the leading-edge vortex system becomes unstable, the amplitude, and the period of the resulting self-sustained limit cycle, in close agreement with two wind-tunnel experiments. With the development of modern aerodynamic configurations employing close-coupled canards, such as the X-29, comes the need to simulate unsteady aerodynamic interference. A versatile method based on the general unsteady vortex-lattice technique has been developed. The method yields the time histories of the pressure distribution on the lifting surfaces, the distribution of vorticity in the wakes, and the position of the wakes simultaneously. As an illustration of the method, the unsteady flowfield for a configuration closely resembling the X-29 is presented. The results show the strong influence of the canards on the main wing, including the time lag between the motions of the canards and the subsequent changes in the vorticity and hence the pressure distributions and loads on the main wing. / Ph. D. / incomplete_metadata

Several procedures for the simultaneous aerodynamic-structural design optimization of aircraft wings are investigated. These procedures include efficient methods for optimization and sensitivity calculations that are applied to two specific design examples. The first is a subsonic transport aircraft with a composite forwardswept wing. The aerodynamic modeling for this case is provided by vortex-lattice theory and the structural model initially utilizes finite-element analyses. Even with efficient sensitivity methods, the approximate optimization problem still requires a large computational effort. To reduce this cost, a variable-complexity model for the structural analyses is introduced. First, an algebraic equation model for wing weight is used in the optimization procedure to obtain an aerodynamic design that approximately accounts for the effects of wing geometry on wing weight. Then this design is refined by simultaneous aerodynamic-structural optimization based on the finite-element analysis. The net effect of this dual structural model is a substantial reduction in optimization costs. The second example is the wing design of a supersonic High-Speed Civil Transport (HSCT). For this case, the simple wing-weight equations for structures are retained. For the aerodynamics, a variable-complexity model was introduced with the complex models provided by volumetric wave drag analysis and panel methods. In addition, simple algebraic models for wave and drag due to lift provide inexpensive approximations during most of the optimization cycles. With the minimization of the costly complex sensitivity calculations, a reduction in optimization costs is realized. / Ph. D.

The rapid air travel growth during the last three decades, has resulted in runway congestion at major airports. The current airports infrastructure will not be able to support the rapid growth trends expected in the next decade. Changes or upgrades in infrastructure alone would not be able to satisfy the growth requirements, and new airplane concepts such as the NASA proposed Super Short Takeo and Landing and Extremely Short Takeo and Landing (ESTOL) are being vigorously pursued. Aircraft noise pollution during Takeoff and Landing is another serious concern and efforts are aimed to reduce the airframe noise produced by Conventional High Lift Devices during Takeoff and Landing. Circulation control technology has the prospect of being a good alternative to resolve both the aforesaid issues. Circulation control airfoils are not only capable of producing very high values of lift (Cl values in excess of 8.0) at zero degree angle of attack, but also eliminate the noise generated by the conventional high lift devices and their associated weight penalty as well as their complex operation and storage. This will ensure not only satisfying the small takeoff and landing distances, but minimal acoustic signature in accordance with FAA requirements. The Circulation Control relies on the tendency of an emanating wall jet to independently control the circulation and lift on an airfoil. Unlike, conventional airfoil where rear stagnation point is located at the sharp trailing edge, circulation control airfoils possess a round trailing edge, therefore the rear stagnation point is free to move. The location of rear stagnation point is controlled by the blown jet momentum. This provides a secondary control in the form of jet momentum with which the lift generated can be controlled rather the only available control of incidence (angle of attack) in case of conventional airfoils. The use of Circulation control despite its promising potential has been limited only to research applications due to the lack of a simple prediction capability. This research effort was focused on the creation of a rapid prediction capability of Circulation Control Aerodynamic Characteristics which could help designers with rapid performance estimates for design space exploration. A morphological matrix was created with the available set of options which could be chosen to create this prediction capability starting with purely analytical physics based modeling to high fidelity CFD codes. Based on the available constraints, and desired accuracy metamodels has been created around the two dimensional circulation control performance results computed using Navier Stokes Equations (Computational Fluid Dynamics). DSS2, a two dimensional RANS code written by Professor Lakshmi Sankar was utilized for circulation control airfoil characteristics. The CFD code was first applied to the NCCR 1510-7607N airfoil to validate the model with available experimental results. It was then applied to compute the results of a fractional factorial design of experiments array. Metamodels were formulated using the neural networks to the results obtained from the Design of Experiments. Additional validation runs were performed to validate the model predictions. Metamodels are not only capable of rapid performance prediction, but also help generate the relation trends of response matrices with control variables and capture the complex interactions between control variables. Quantitative as well as qualitative assessments of results were performed by computation of aerodynamic forces and moments and flow field visualizations. Wing characteristics in three dimensions were obtained by integration over the whole wing using Prandtl's Wing Theory. The baseline Super STOL configuration was then analyzed with the application of circulation control technology. The desired values of lift and drag to achieve the target values of Takeoff and Landing performance were compared with the optimal configurations obtained by the model. The same optimal configurations were then subjected to Super STOL cruise conditions to perform a tradeoff analysis between Takeoff and Cruise Performance. Supercritical airfoils modified for circulation control were also thoroughly analyzed for Takeoff and Cruise performance and may constitute a viable option for Super STOL and STOL Designs. The prediction capability produced by this research effort can be integrated with the current conceptual aircraft modeling and simulation framework. The prediction tool is applicable within the selected ranges of each variable, but methodology and formulation scheme adopted can be applied to any other design space exploration.

Thesis (MScEng) -- Stellenbosch University, 1976. / pt. A. Theory and computer program -- pt. B. Experimental. / INTRODUCTION: The advent of the high speed computer has revolutionized structural design in all spheres of Engineering. Up till then structural stress analysis was limited to over-simplification of the structure in question to comply with derived classical mathematical solutions. In practice however the picture is very different, the structure usually being complex and highly redundant in nature. The techniques involving Energy Methods to solve such structures have been known for a long time. However they required weeks of hand calculations to solve only a small number of redundancies in a structure. Neville Shute mentions this in his book Slide Rule. The development of Matrix algebra and the finite element method has made it possible to analyse, say, a complete aircraft structure in a matter of days. using a large capacity high speed computer~ Experimental results have shown that finite element stress analysis comes much closer to reality than the dated classical methods.

An experimental examination of crossflow instability and transition on a 45° swept wing is conducted in the Arizona State University Unsteady Wind Tunnel. The stationary-vortex pattern and transition location are visualized using both sublimating-chemical and liquid-crystal coatings. Extensive hot-wire measurements are conducted at several measurement stations across a single vortex track. The mean and travelling-wave disturbances are measured simultaneously. Stationary-crossflow disturbance profiles are determined by subtracting either a reference or a span-averaged velocity profile from the mean-velocity data. Mean, Stationary-crossflow, and travelling-wave velocity data are presented as local boundary-layer profiles and as contour plots across a single stationary-crossflow vortex track. Disturbance-mode profiles and growth rates are determined. The experimental data are compared to predictions from linear stability theory. Comparison of measured and predicted pressure distributions shows that a good approximation of infinite swept-wing flow is achieved. A fixed-wavelength vortex pattern is observed throughout the visualization range. The theoretically-predicted maximum-amplified vortex wavelength is found to be approximately 25% larger than the observed wavelength. Linear-stability computations for the dominant stationary-crossflow vortices show that the N-factors at transition ranged from 6.4 to 6.8. The mean-velocity profiles vary slightly across the stationary-crossflow vortex at the first measurement station. The variation across the vortex increases with downstream distance until nearly all of the profiles become highly-distorted S-shaped curves. Local stationary-crossflow disturbance profiles having either purely excess or deficit values develop at the upstream measurement stations. Further downstream the profiles take on crossover shapes not anticipated by the linear theory. The maximum streamwise stationary-crossflow velocity disturbances reach +20% of the edge velocity just before transition. The travelling-wave disturbances have single lobes at the upstream measurement stations as expected, but further downstream double-lobed travelling-wave profiles develop. The maximum disturbance intensity remains quite low until just ahead of the transition location where it suddenly peaks at 0.7% of the edge velocity and then drops sharply. The travelling-wave intensity is always more than an order of magnitude lower than the stationary crossflow-vortex strength. The mean streamwise-velocity contours are nearly flat and parallel to the model surface at the first measurement station. Further downstream, the contours rise up and begin to roll over like a wave breaking on the beach. The stationary-crossflow contours show that a plume of low-velocity fluid rises near the center of the wavelength while high-velocity regions develop near the surface at each end of the wavelength. There is no distinct pattern to the low-intensity travelling-wave contours until a short distance upstream of the transition location where the travelling-wave intensity suddenly peaks near the center of the vortex and then falls abruptly. The experimental disturbance-mode profiles agree quite well with the predicted eigenfunctions for the forward measurement stations. At the later stations, the experimental mode profiles assume double-lobed shapes with maxima above and below the single maximum predicted by the linear theory. The experimental growth rates are found to be less than or equal to the predicted growth rates from the linear theory. Also, the experimental growth rate curve oscillates over the measurement range whereas the theoretically-predicted growth rates decrease monotonically. / Ph. D.

Subsonic wind tunnel investigations were conducted on a 60° swept, flat plate, delta wing with a leading edge vortex flap. The pressure distributions were measured over a range of angles of attack starting from zero to 40° in 5° interval and flap deflection angles from zero to 45° with 5° increments at a Reynolds number of about 2.14 x 10‘ based on the root chord. The flow visualization experiments were performed from zero degree to the stall angle, with ten different flap deflection angles at the same Reynolds number. The mean flow field was measured at angles of attack l0° and 15° with the flap deflection angles of l0° and 30° at a Reynolds number of about 1.50 x 10°. The experimental results shows that the leading edge vortex flap is an effective means to control the vortex flow over a delta wing. The optimum flap deflection angles were found where the primary vortex was confined to the leading edge vortex flap, thus producing a thrust on the flap. It was found that flap deflection could be used to restore a vortex flow from burst vortex condition. / Ph. D. / incomplete_metadata

Low aspect ratio, highly-swept cranked arrow wing planforms are often proposed for high-speed civil transports. These wing planforms offer low supersonic drag without suffering greatly from low lift/drag ratios in low-speed flight. They can, however, suffer from pitch-up at modest angles of attack (as low as 5°) during low-speed flight due to leading edge vortex influence, flow separation and vortex breakdown. The work presented here describes an investigation conducted to study past research on the longitudinal aerodynamic characteristics of highly-swept cranked wing planforms, the development of a new method to estimate pitch-up of these configurations, and the applications of this new method to the analysis of tail designs for trim at high lift coefficients. The survey of past research placed emphasis on 1) understanding the problem of pitch-up, 2) ascertaining the effects of leading and trailing edge flaps, and 3) determining the benefits and shortfalls of tail, tailless, and canard configurations. The estimation method used a vortex lattice method to calculate the inviscid flow solution. Then, the results were adjusted to account for flow separation on the outboard wing section by imposing a limit on the equivalent 2-D sectional lift coefficient. The new method offered a means of making low cost estimates of the nonlinear pitching moment characteristics of slender, cranked arrow wing configurations with increased accuracy compared to conventional linear methods. Numerous comparisons with data are included. The new method was applied to analyze the trim requirement of slender wing designs generated by an aircraft configuration optimization and design program. The effects of trailing edge flaps and horizontal tail on the trimmed lift coefficient was demonstrated. Finally, recommendations were made to the application of this new method to multidisciplinary design optimization methods. / Master of Science