Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2009. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3623. Adviser: Eric Loth. Includes bibliographical references (leaves 160-165) Available on microfilm from Pro Quest Information and Learning.

Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2009. / Source: Dissertation Abstracts International, Volume: 70-06, Section: B, page: 3755. Adviser: Jonathan Freund. Includes bibliographical references (leaves 107-110) Available on microfilm from Pro Quest Information and Learning.

<p>Computer simulations of complex mathematical models are a powerful tool for design, but they introduce uncertainties which can lead to poor design choices when simulation data is all that is available. Additionally, computational grid generation can dramatically increases the costs associated with initializing numerical simulations. Proper verification can help quantify the uncertainty in numerical simulations, and a new form of code verification is presented. This is based on the method of manufactured solutions for integral equations, which allows MMS to be used to verify shock-capturing codes. A procedure is presented for numerically evaluating the required integrals, and it is found to completely eliminate numerical error resulting from discontinuous integrand functions. Integral MMS is demonstrated, and it is found to yield convergence rates that differ by less than 5% from those obtained using differential MMS, and which match precisely with the theoretical rates for discontinuous solutions. This indicates that integral MMS can be used for code verification in place of differential MMS, which cannot be used with discontinuous solutions. Moving grids can be used to allow computed fluid motion to generate the computational grid automatically. The unique challenges associated with grid motion are explored, and multiple implementations are discussed. A software library for fluid-mechanical simulation in unsteady coordinates is also introduced. Preliminary verification of both the method and the library is discussed. The use of unsteady coordinates affects accuracy and grid convergence rates in complex ways. This work lays the foundation for future work on the use of moving grids in order to reduce the grid-generation burden for design-oriented computational fluid dynamics.

<p> A comprehensive survey of available numerical methods and models was performed on the open source computational fluid dynamics solver OpenFOAM version 2.0 for incompressible turbulent bluff body flows. Numerical methods are illuminated using source code for side-by-side comparison. For validation, the accuracy of flow predictions over a sphere in the subcritical regime and delta wing with sharp leading edge is assessed. Solutions show mostly good agreement with experimental data and data obtained from commercial software. A demonstration of the numerical implementation of a dynamic hybrid RANS/LES framework is also presented, including results from test studies. </p>

<p> The prolific development of Unmanned Aerial Vehicles (UAV&rsquo;s) in recent years has the potential to provide tremendous advantages in military, commercial and law enforcement applications. While safety and performance take precedence in the development lifecycle, autonomous operations and, in particular, cooperative missions have the ability to significantly enhance the usability of these vehicles. The success of cooperative missions relies on the optimal allocation of targets while taking into consideration the resource limitation of each vehicle. The task allocation process can be centralized or decentralized. This effort presents the development of a real time target allocation algorithm that considers available stored energy in each vehicle while minimizing the communication between each UAV. The algorithm utilizes a nearest neighbor search algorithm to locate new targets with respect to existing targets. Simulations show that this novel algorithm compares favorably to the mixed integer linear programming method, which is computationally more expensive. The implementation of this algorithm on Arduino and Xbee wireless modules shows the capability of the algorithm to execute efficiently on hardware with minimum computation complexity.</p>

A new approach is developed for the analysis and design of mechanically fastened composite lap joints under mechanical and thermal loading. Based on the combined complex potential and variational formulation, the solution method satisfies the equilibrium equations exactly while the boundary conditions are satisfied by minimizing the total potential. This approach is capable of modeling finite laminate planform dimensions, uniform and variable laminate thickness, laminate lay-up, interaction among bolts, bolt torque, bolt flexibility, bolt size, bolt-hole clearance and interference, insert dimensions and insert material properties. Comparing to the finite element analysis, the robustness of the method does not decrease when modeling the interaction of many bolts; also, the method is more suitable for parametric study and design optimization. The Genetic Algorithm (GA), a powerful optimization technique for multiple extrema functions in multiple dimensions search spaces, is applied in conjunction with the complex potential and variational formulation to achieve optimum designs of bolted composite lap joints. The objective of the optimization is to acquire such a design that ensures the highest strength of the joint. The fitness function for the GA optimization is based on the average stress failure criterion predicting net-section, shear-out, and bearing failure modes in bolted lap joints. The criterion accounts for the stress distribution in the thickness direction at the bolt location by applying an approach utilizing a beam on an elastic foundation formulation.

Direct Numerical Simulations (DMS) of the incompressible Navier-Stokes equations are used to investigate the effect of wall suction on transition in a flat-plate boundary layer. The Navier-Stokes equations are cast in vorticity-velocity formulation. The streamwise and wall-normal derivatives are discretized with compact differences, with a pseudospectral treatment of the spanwise derivatives. Two different methods are used for the time integration. In most calculations, an explicit four-stage Runge Kutta method is used. In some cases, a semi-implicit combination of a three-stage Runge-Kutta- and a Crank-Nicolson method is used. Several case studies are performed. The first case treats the effect of a single row of suction holes, aligned in the spanwise direction, on the evolution of a Tollmien-Schlichting wave. It is found that suction through small holes leads to noticeable nonlinear effects on disturbances with large spanwise wavenumbers. The effect of suction on secondary instability with regards to a large-amplitude Tollmien-Schlichting wave is investigated in the second case study. The suction configurations here are a permeable wall, spanwise slots, and streamwise slots. It is found that sufficiently strong suction suppresses the secondary instability. The different suction configurations are equally effective. The role of the Klebanoff-mode in boundary layer transition is the subject of the third case study. A numerical model of the Klebanoff-mode is presented that agrees well with experimental observations. It is shown how the interaction between the Klebanoff-mode and a Tollmien-Schlichting wave can cause transition. Wall suction is found to be an effective means to prevent transition and maintain laminar flow even in the presence of high-amplitude Klebanoff-mode fluctuations. In the last case study, the limit of very strong suction through holes is investigated. It is shown how the suction holes generate streamwise vortices that can become unstable and lead to bypass transition.

This study presents a new straight pipe element that enables the efficient computation of large, three-dimensional deformations in pipes with circular cross-sections. Existing pipe elements have difficulties in applications including three-dimensional large deformations. To overcome these diffculties, the new element (called RC pipe element in this work), which supports rigid-body and constant-strain modes, is modeled using curvilinear shell coordinates and sinusoidal interpolation functions. This new element captures all stresses except the normal stress across the shell thickness (i.e. small thickness is assumed). Euler parameters are used to describe rotational rigid-body modes and are incorporated into the element's vector of degrees of freedom. Under general loading (axial, transverse, bending and torsion), the element allows large ovalization of its cross section and large, three-dimensional angular changes in the orientation of its reference axis. The formulation used to derive the element incorporates the nonlinear coupling between torsional and bending deformations. Results of the static analysis are presented for the stresses and deformations produced by combined bending and torsional loads. A comparison of these results to corresponding quantities generated by ABAQUS using a large number of 24 degree-of-freedom shell elements indicates excellent agreement and significant gains in computational efficiency because of a reduction in number of degrees of freedom. Results are also presented for the behaviour of the element in the presence of large rotational motion with internal pressure.

Mechanical fasteners remain the primary means of load transfer between structural components made of composite laminates. As, in pursuit of increasing efficiency of the structure, the operational load continues to grow, the load carried by each fastener increases accordingly. This accelerates initiation of fatigue-related cracks near the fasteners holes and increases probability of failure. Therefore, the assessment of the stresses around the fastener holes and the stress intensity factors associated with edge cracks becomes critical for damage-tolerant design. Because of the presence of unknown contact stresses and the contact region between the fastener and the laminate, the analysis of a pin-loaded hole becomes considerably more complex than that of a traction-free hole. The accurate prediction of the contact stress distribution along the hole boundary is critical for determining the stress intensity factors and is essential for reliable strength evaluation and failure prediction. This study concerns the development of an analytical methodology, based on the boundary collocation technique, to determine the contact stresses and stress intensity factors required for strength and life prediction of bolted joints with many fasteners. It provides an analytical capability for determining the non-linear contact stresses in mechanically fastened composite laminates while capturing the effects of finite geometry, presence of edge cracks, interaction among fasteners, material anisotropy, fastener flexibility, fastener-hole clearance, friction between the pin and the laminate, and by-pass loading. Also, the proposed approach permits the determination of the fastener load distribution, which significantly influences the failure load of a multi-fastener joint. The well known phenomenon of the fastener tightening torque (clamping force) influence on the load distribution among the different fastener in a multi-fastener joints is taken into account by means of bi-linear representation of the elastic fastener deflection. Finally, two different failure criteria, maximum strains averaged over the characteristic distances and Tsai-Wu criterion, were used to predict the failure load and failure mode in two composite-aluminum joints. The comparison of the present predictions with the published experimental results reveals their agreement.

An experiment was designed and built to study vertical annular air-water flows in a channel with a rectangular cross section with no heat transfer. Flush-wire electrical conductivity probes were theoretically analyzed to demonstrate their potential to accurately measure liquid film thickness in the experiment with high temporal and spatial resolution. Flush-wire probes were then successfully implemented and film thickness measurements obtained. From theoretical analysis, the suitability of micromachined hot film and floating element wall shear stress sensors for measurements of wall shear stress in the annular flow was investigated. A microfabricated hot film wall shear stress sensor was subsequently packaged and installed in the experiment, where it provided wall shear stress measurements with high temporal and spatial resolution. After the implementation of these new measurement techniques, a large suite of test cases was run and data for film thickness and wall shear stress acquired. A statistical analysis of the film thickness data indicates the existence of two distinct wave regimes, ripple waves and disturbance waves, within the annular flow regime. Spectral decomposition of film thickness and wall shear stress data demonstrates the existence of dominant frequencies in the wave spectrum and an exponential decay of wave amplitudes at high frequencies indicative of a force balance between capillary and momentum forces. Wave velocities were determined from cross correlations which again provided evidence of different types of waves each with different wave velocities and spatial extensions. A semi-analytical model for wave velocities as a function of superficial Reynolds numbers was validated and improved. The improved model gives an accurate prediction for wave velocities and is based on physical arguments representing the appropriate length scales in annular flow. The experimental results and data analysis provide a new perspective of annular two-phase flows in a channel with a rectangular cross section.