Master of Science (M.S.), University of Dayton, 2018, Aerospace Engineering

This work seeks to provide a proof-of-concept for the use of variable-fidelity (VF) kriging to approximate the lift and drag values for a complex hypersonic flight vehicle. Otherwise known as aerodynamic database generation within the aerospace engineering community, the force or moment experienced by a vehicle due to airflow, as a function of independent inputs such as flight speed or attitude, is approximated via some mathematical form. In the case of this work, VF kriging is implemented such that the vehicle response is interpolated directly through the points of high-fidelity (HF) simulation data while the trends of the response approximation are guided by low-fidelity (LF) information. High-fidelity simulations are implemented via the Euler flow computational software package Cart3D. The low-fidelity information is given by supersonic-hypersonic small-disturbance theory implemented in a surface pressure estimation code, developed specifically for this work for completely arbitrary body shapes represented by unstructured, triangular-cell surface meshes. The major contribution is a framework that connects the two fidelity levels to VF kriging routines to produce lift and drag approximations of arbitrary complex vehicles under hypersonic flight conditions. Assessment of the quality of the approximations is given by the root-mean-square error (RMSE) between the VF kriging surrogates and high-fidelity simulations performed over the same independent input domain. Results in two dimensions show that the use of VF kriging, to produce an interpolant as a function of angle-of-attack and Mach number, increases surrogate accuracy by nearly an order of magnitude for lift and by over twenty times for drag, when compared to ordinary kriging without variable-fidelity modeling. Three-dimensional surrogates, with input of angle-of-attack and two independent elevon control surface deflections, show roughly two and four times more accuracy for lift and drag, respectively, compared (open full item for complete abstract)

Committee: Markus Rumpfkeil (Advisor); Jose Camberos (Committee Member); Raymond Kolonay (Committee Member) Subjects: Aerospace Engineering; Applied Mathematics

Doctor of Philosophy, The Ohio State University, 2013, Aero/Astro Engineering

This dissertation describes model reduction techniques for the computation of aerodynamic heat flux and pressure loads for multi-disciplinary analysis of hypersonic vehicles. NASA and the Department of Defense have expressed renewed interest in the development of responsive, reusable hypersonic cruise vehicles capable of sustained high-speed flight and access to space. However, an extensive set of technical challenges have obstructed the development of such vehicles. These technical challenges are partially due to both the inability to accurately test scaled vehicles in wind tunnels and to the time intensive nature of high-fidelity computational modeling, particularly for the fluid using Computational Fluid Dynamics (CFD). The aim of this dissertation is to develop efficient and accurate models for the aerodynamic heat flux and pressure loads to replace the need for computationally expensive, high-fidelity CFD during coupled analysis. Furthermore, aerodynamic heating and pressure loads are systematically evaluated for a number of different operating conditions, including: simple two-dimensional flow over flat surfaces up to three-dimensional flows over deformed surfaces with shock-shock interaction and shock-boundary layer interaction. An additional focus of this dissertation is on the implementation and computation of results using the developed aerodynamic heating and pressure models in complex fluid-thermal-structural simulations. Model reduction is achieved using a two-pronged approach. One prong focuses on developing analytical corrections to isothermal, steady-state CFD flow solutions in order to capture flow effects associated with transient spatially-varying surface temperatures and surface pressures (e.g., surface deformation, surface vibration, shock impingements, etc.). The second prong is focused on minimizing the computational expense of computing the steady-state CFD solutions by developing an efficient surrogate CFD model. The develop (open full item for complete abstract)

Committee: Jack McNamara (Advisor); Thomas Eason III (Committee Member); Jeffrey Bons (Committee Member); Mo-How Herman Shen (Committee Member); Mei Zhuang (Committee Member) Subjects: Aerospace Engineering

Doctor of Philosophy, The Ohio State University, 2023, Electrical and Computer Engineering

We present a discussion on the reduced-order modeling of electromagnetic simulation in general, and kinetic plasma simulations in particular, using the Proper Orthogonal Decomposition technique. Computational electromagnetics has been an important tool for physicists and engineers since the mid-1960s, when the increasing availability of modern high-speed computers started to allow the numerical solution of practical problems for which closed-form analytic solutions did not exist or were impractical to calculate. The study of kinetic plasmas is of great interest both for theoretical exploration and technological applications such as design of vacuum electronic devices, the study of the interaction of space-borne assets and cosmic radiation, fusion experiments, among others. Due to the theoretical complexity of these problems and the difficulty in performing physical experiments, simulations are instrumental for obtaining new insights or developing new device designs by resolving the field and plasma behaviors when changes are made. Several variants of simulations exist, but particle-in-cell algorithms for solving particle dynamics coupled with finite-differences or finite-elements field solvers are particularly successful. Despite their success, such algorithms are still constrained by computational cost such as processing time and memory/storage limitations. The Proper Orthogonal Decomposition is a technique that extracts the spatiotemporal behavior from a function of interest or a set of data points. This spatiotemporal behavior is characterized by a set of coupled spatial and temporal modes, which makes the Proper Orthogonal Decomposition especially suitable for analyses and applications in dynamic systems; it has been used for creation of reduced-order models in the past, especially in the fluid dynamics community where it originated from but also in many other areas. We have explored the application of the Proper Orthogonal Decomposition technique to co (open full item for complete abstract)

Committee: Fernando Teixeira (Advisor); Casey Wade (Committee Member); Kubilay Sertel (Committee Member); Robert Burkholder (Committee Member) Subjects: Electrical Engineering; Electromagnetics

Doctor of Philosophy, The Ohio State University, 2021, Aerospace Engineering

Balancing accurate and efficient estimation of aerothermodynamic loads is a significant challenge for multi-disciplinary modeling and analysis of high speed vehicles. High fidelity methods are desired in order to minimize modeling uncertainty. However, the need for either online aerothermodynamic modeling or many model iterations often introduces a hard constraint on model run-time that favors the use of classical engineering methods. Thus, systematic characterization of trade-offs in accuracy and run-time costs for different levels of modeling fidelity is a critical need. Technical challenges toward this need include the wide range of operating conditions and parameters associated with a complete vehicle, as well as the inability to comprehensively assess hypersonic configurations in ground-based facilities. The goal of this dissertation is to systematically study a range of reduced fidelity aerothermodynamics models in order to determine relevant fidelity. This is carried out by first examining the impact of analytical and data-driven model reductions over a broad parameter space in the context of steady-state aerothermodynamic loads. This work builds on previous by extending existing approaches to accommodate a complete three-dimensional hypersonic geometry. Three models for pressure are investigated: two of the models are kriging interpolants of either Reynolds-averaged Navier-Stokes (RANS) or Euler computational fluid dynamics (CFD) data, and the third is a shock-expansion model. Four models for heat flux are investigated: a kriging interpolant of RANS CFD; an Eckert's reference temperature model using a kriging interpolant of RANS CFD pressure; an Eckert's reference temperature model using kriging interpolants of Euler CFD pressure and temperature; and an Eckert's reference temperature model using shock-expansion theory pressure and temperature. The reduced models are compared to RANS CFD predictions on the surface of a complete hypersonic vehicle over a b (open full item for complete abstract)

Committee: Jack McNamara Professor (Advisor); Carlos E.S. Cesnik Professor (Committee Member); Jen-Ping Chen Professor (Committee Member); Reasor Reasor Ph.D. (Committee Member); Mei Zhuang Professor (Committee Member); Abdollah Shafieezadeh Professor (Committee Member) Subjects: Aerospace Engineering; Engineering; Mechanical Engineering; Statistics

Doctor of Philosophy, The Ohio State University, 2020, Mechanical Engineering

Understanding and utilizing the dynamics of piecewise-linear (PWL) nonlinear systems are of crucial important for many engineered systems. These systems include complex mechanical/aerospace systems where intermittent contact significantly affects performance such as cracked/damaged structures and nonlinear energy harvesters. The research in this dissertation is focused on the development of efficient computational tools for analyzing and controlling the nonlinear dynamics of these systems. Current techniques are not able to effectively predict the full dynamics of complex PWL nonlinear systems due to the lack of computational efficiency and limited capability of current methods, and hence obstruct the development of new technologies. This dissertation has two main goals. First, a new class of effective techniques that enable the computation of the response of large PWL nonlinear systems is developed to analyze, monitor and control their dynamics. These methods utilize linear features in the PWL nonlinear systems and integrate them with numerical optimization tools to capture the nonlinear response of these systems. Both transient and steady-state dynamic responses can be analyzed with several orders of magnitudes speed-up compared to current techniques. The second goal of this dissertation is the application of the general methods in two engineering areas: (1) modeling the nonlinear dynamics of damaged/cracked structures and (2) the active control for optimized vibration performance of energy harvesters. The dynamic behavior of cracked engineering structures are investigated to advance the understanding of structural damage. A better understanding of changes in dynamics caused by structural damage could advance the development of damage monitoring techniques that prevent catastrophic failures of many systems. Also, the proposed techniques provides an efficient way to control the vibration of PWL nonlinear systems so that a more effective extraction of energy from (open full item for complete abstract)

Committee: Kiran D'Souza (Advisor); Manoj Srinivasan (Committee Member); Ahmet Kahraman (Committee Member); M.-H. Herman Shen (Committee Member) Subjects: Mechanical Engineering

Doctor of Philosophy, University of Akron, 2019, Mechanical Engineering

Mechanical connection of parts through jointed connections are prolific throughout modern engineering applications; however, precision analysis and design of these systems remains difficult. Experimental findings have revealed a myriad of nonlinear properties of these systems such as nonlinear damping, hysteresis, etc., and these complex effects lead to extreme difficulties in the characterization and modeling of these common structural elements. To exacerbate matters, high-fidelity numerical analysis of these systems is often impractical due to disparate length and time-scales between microslip in the joint and macro-scale effects of interest. In this dissertation, original research on the analysis of damping of jointed structures is presented. This includes theoretical work in advancement of reduced-order modal models as well as practical development of Abaqus subroutines to implement cutting-edge damping models into finite element models. This work culminates in the study of a practical problem of interest to Sandia National Labs involving a jointed structure under blast loading, and important conclusions are draw about the nature of jointed structures under complex loads.

Committee: Donald Quinn (Advisor); Graham Kelly (Committee Member); Xiaosheng Gao (Committee Member); Ernian Pan (Committee Member); Kevin Kreider (Committee Member) Subjects: Aerospace Engineering; Applied Mathematics; Mathematics; Mechanical Engineering; Mechanics

Doctor of Philosophy, The Ohio State University, 2016, Aero/Astro Engineering

Galerkin projection is a commonly used reduced order modeling approach; however, stability and accuracy of the resulting reduced order models are highly dependent on the modal decomposition technique used. In particular, deriving stable and accurate reduced order models from highly turbulent flow fields is challenging due to the presence of multi-scale phenomenon that cannot be ignored and are not well captured using the ubiquitous Proper Orthogonal Decomposition (POD). A truncated set of proper orthogonal modes is biased towards energy dominant, large-scale structures and results in over-prediction of kinetic energy from the corresponding reduced order model. The accumulation of energy during time-integration of a reduced order model may even cause instabilities. A modal decomposition technique that captures both the energy dominant structures and energy dissipating small scale structures is desired in order to achieve a power balance. The goal of this dissertation is to address the stability and accuracy issues by developing and examining alternative basis identification techniques. In particular, two modal decomposition methods are explored namely, sparse coding and Dynamic Mode Decomposition (DMD). Compared to Proper Orthogonal Decomposition, which seeks to truncate the basis spanning an observed data set into a small set of dominant modes, sparse coding is used to identify a compact representation that span all scales of the observed data. Dynamic mode decomposition seeks to identity bases that capture the underlying dynamics of a full order system. Each of the modal decomposition techniques (POD, Sparse, and DMD) are demonstrated for two canonical problems of an incompressible flow inside a two-dimensional lid-driven cavity and past a stationary cylinder. The constructed reduced order models are compared against the high-fidelity solutions. The sparse coding based reduced order models were found to outperform those developed using the dynamic mode and (open full item for complete abstract)

Committee: Jack McNamara (Advisor); Datta Gaitonde (Committee Member); Ryan Gosse (Committee Member); Joseph Hollkamp (Committee Member); Mohammad Samimy (Committee Member) Subjects: Aerospace Engineering

Master of Science in Engineering, University of Akron, 2016, Mechanical Engineering

Electric motors are an extremely important class of electro-mechanical machines and are used today in countless applications. In particular, motors with distinctive poles, e.g. switched reluctance motors, see considerable use in both industry and academia; however, this class of motors suffers from acoustic and vibration issues as a result of the pulsating nature of its torque production. To make matters worse, analytical and numerical modeling of many aspects of these machines is extremely difficult due to their nonlinear, electromagnetic-mechanically coupled nature. Finite element methods (FEM) can provide some relief to designers, but at great computational costs. This thesis models, numerically simulates, and studies these electric motors using first principles so as to provide future designers with a physics-based approach analyzing the mechanical vibrations of these systems. Two varieties of machines are analyzed: switched reluctance motors (SRMs) and modular transverse flux motors (MTFMs), specifically a novel MTFM developed by colleagues from the electrical engineering field. A modular methodology is developed to suit a variety of design cases, and both vibration and stress analyses are conducted to show a variety of post-processing results.

Committee: Donald Quinn PhD (Advisor); Yilmaz Sozer PhD (Committee Member); Graham Kelly PhD (Committee Member); Gerald Young PhD (Committee Member) Subjects: Mechanical Engineering

Doctor of Philosophy (PhD), Wright State University, 2008, Engineering PhD

Design of structural components is constrained by both iteration time and prediction uncertainty. Iteration time refers to the computation time each simulation requires and controls how much design space can be explored given a fixed period. A comprehensive search of the space leads to more optimum designs. Prediction uncertainty refers to both irreducible uncertainties, such as those caused by material scatter, and reducible uncertainty, such as physics-based model error. In the presence of uncertainty, conservative safety factors and design margins are used to ensure reliability, but these negatively impact component weight and design life. This research investigates three areas to improve both iteration time and prediction uncertainty for turbomachinery design. The first develops an error-quantified reduced-order model that predicts the effect of geometric deviations on airfoil forced response. This error-quantified approximation shows significant improvements in accuracy compared to existing methods because of its bias correction and description of random error. The second research area develops a Probabilistic Gradient Kriging approach to efficiently model the uncertainty in predicted failure probability caused by small sample statistics. It is shown that the Probabilistic Gradient Kriging approach is significantly more accurate, given a fixed number of training points, compared to conventional Kriging and polynomial regression approaches. It is found that statistical uncertainty from small sample sizes leads to orders of magnitude variation in predicted failure probabilities. The third research area develops non-nominal and nominal mode Component Mode Synthesis methods for reduced-order modeling of the geometric effects on rotor mistuning. Existing reduced-order methods approximate mistuning with a nominal-mode, or design intent, basis and airfoil modal stiffness perturbation. This assumption introduces error that can be quantified when compared to a finite el (open full item for complete abstract)

Committee: Ramana Grandhi PhD (Advisor); Joseph Slater PhD (Committee Member); Ravi Penmetsa PhD (Committee Member); Mo-how Shen PhD (Committee Member); Charles Cross PhD (Committee Member) Subjects: Mechanical Engineering

PhD, University of Cincinnati, 2006, Engineering : Mechanical Engineering

Constructing nonlinear structural dynamic models, useful for sonic fatigue prediction purposes, has been a goal of the United States Air Force (USAF) for decades. Such a predictive capability is required in the development of advanced, high-performance aircraft structures. Specifically, the USAF is seeking the ability to predict the response of complex structures to engine induced and aero-acoustic loading. Sonic fatigue has plagued the USAF since the advent and adoption of the turbine engine. While the problem has historically been a maintenance one, predicting the dynamic response is crucial for future aerospace vehicles. Decades were spent investigating the dynamic response and untimely failure of aircraft structures, yet little work was accomplished towards developing practical nonlinear prediction methods. Further, the last decade witnessed an appreciable amount of work in the area of nonlinear parameter identification. This study outlines and validates a unique and important extension of a recently introduced nonlinear identification method; Nonlinear Identification through Feedback of the Outputs (NIFO). The novel extension allows for a ready means of identifying nonlinear parameters in reduced order space using experimental data. The nonlinear parameters are then used in the assembly of reduced order models thus providing researchers with a means of conducting predictive studies prior to expensive and questionable experimental efforts. This research details both an analytical and experimental study conducted on a well characterized clamped-clamped beam subjected to broadband random loading. Amplitude dependent, constant stiffness parameters were successfully identified for both single and multiple degree-of-freedom (SDOF, MDOF) nonlinear reduced order models. The nonlinear coefficients identified from the analytical scenario compare well with previously published studies of the beam. Nonlinear parameters were also successfully identified from the raw experim (open full item for complete abstract)

Committee: Dr. Randall Allemang (Advisor) Subjects: Engineering, Mechanical

Doctor of Philosophy, The Ohio State University, 2007, Electrical Engineering

Flow control refers to the ability to manipulate fluid flow so as to achieve a desired change in its behavior, which offers many potential technological benefits, such as reducing fuel costs for vehicles and improving effectiveness of industrial processes. An interesting case of flow control is cavity flow control, which has been the motivation of this study: When air flow passes over a shallow cavity a strong resonance is produced by a natural feedback mechanism, scattering acoustic waves that propagate upstream and reach the shear layer, and developing flow structures. These cause many practical problems including damage and fatigue in landing gears and weapons bays in aircrafts. Presently there is a lack of sufficient mathematical analysis and control design tools for flow control problems. This includes mathematical models that are amenable to control design. Recently reduced-order modeling techniques, such as those based on proper orthogonal decomposition (POD) and Galerkin projection (GP), have come to interest. However, a main issue with these models is that the effect of boundary conditions, which is where the control input is, gets embedded into system coefficients. This results in a form quite different from what one deals with in standard control systems framework, which is a set of ordinary differential equations (ODE) where the input appears as an explicit term. Another issue with the standard POD/GP models is that they do not yield to systems that have any apparent structure in their coefficients. This leaves one with little choice other than to neglect the nonlinearities of the models and employ standard linear control theory based designs. The research presented in this thesis makes an effort at closing the gaps mentioned above by 1) presenting a reduced-order modeling method utilizing a novel technique for input separation on POD/GP models, 2) introducing a technique based on averaging theory and center manifold theory so as to reveal certain struct (open full item for complete abstract)

Committee: Andrea Serrani (Advisor) Subjects: