Dissertations and Theses - Aerospace Engineering
http://hdl.handle.net/2142/14800
Wed, 25 Feb 2015 08:55:14 GMT2015-02-25T08:55:14ZThe role of molecular length scale on the viscoplastic response of amorphous polymer nanofibers in the glassy state
http://hdl.handle.net/2142/73086
The role of molecular length scale on the viscoplastic response of amorphous polymer nanofibers in the glassy state
Polymeric nanostructures are quite distinct because of the large ratio of surface-to-volume macromolecules which result in very different physical and mechanical behavior compared to bulk. Of special interest are nanostructures in which the constituent polymer is in its glassy state to provide structural and dimensional stability against surface forces that are particularly strong at the micron and submicron scales. The majority of existing literature on the mechanical properties of ultra-small volumes of polymers in their glassy state focuses on ultra-thin films and is limited to small deformations and the viscoelastic regime. Experiments with composite materials encapsulating polystyrene (PS) thin films and in their glassy state at room temperature have demonstrated shear yielding and large deformations that are not possible at the macroscale where PS fails at small strains due to crazing. This dissertation research focused on direct experiments with atactic PS nanofibers to elucidate and quantify the unusual viscoplastic response of ultra-small volumes of PS as a function of the underlying molecular and structural length scales. PS is an ideal polymer for this study because it is amorphous and its glass transition temperature, Tg, is much higher (100C) than room temperature.
The objective of this research was to understand the synergistic coupling between the material length scale as defined by macromolecular size, and the specimen size as defined by the fiber diameter, which can result in extreme ductility and simultaneous strengthening and toughening for fiber diameters at the submicron scale. To this goal, PS fibers with diameters 150–5,000 nm were electrospun from monodisperse PS powders with molecular weights, MW, in the range 13,000–9,000,000 g/mol. Individual nanofibers were tested using a surface micromachined device for nanofiber testing at the quasi-static strain rate of 10-2 s-1. Unlike the brittle behavior of bulk PS, the engineering stress vs. stretch ratio of individual nanofibers with several combinations of MW and diameter displayed very repeatable post-yield behavior including necking and pronounced strain-hardening. Specifically, the ratio of the structural length scale (fiber diameter) to the intrinsic macromolecular length scale (root-mean-square end-to-end chain distance), Dnorm, was shown be an excellent scaling parameter to determine the occurrence and evolution of necking and strain hardening in submicron scale PS fibers. This interplay between molecular and structural length scales in glassy PS fibers was favorably exploited to harness a ~3,000% increase in toughness along with simultaneous increase in tensile strength: the highest fiber strength was achieved for Dnorm = 3–5, whereas increasing Dnorm resulted in gradual decline in strength. Bulk-like brittle behavior took place for Dnorm > 18. It was shown that the effects of molecular and structural lengths scales on large deformation behavior of fibers could be collapsed onto a single master curve as long as the MW was larger than the critical value for constant inter-chain entanglement length. Furthermore, it was shown that the pronounced hardening in PS nanofibers is not a result of post yield necking, but part of the material constitutive response: experiments on individual poly(lactide-co-glycolic acid) (PLGA) nanofibers showed that unlike in bulk specimens, nanoscale imperfections and specimen irregularities are rather benign and strong post-yield strain hardening occurs even when necking is suppressed.
The viscous component of the large deformation response in PS nanofibers was assessed by tensile experiments with PS nanofibers with MW = 123,000 – 2,000,0000 g/mol and diameters of 200 – 750 nm in the range of strain rates 10-4 - 102 s-1. For fibers with Dnorm < 8, it was shown that increasing strain rate resulted in monotonic increase of the stress amplitude without affecting the large fiber stretch ratios. In contrast, the strain rate influenced both the stress and the stretch ratio of fibers with Dnorm > 10, i.e. fibers without significant post-neck hardening. For all PS fibers, the rate dependent stress vs. stretch ratio curves scaled with yield stress. Therefore, a normalized stress vs. strain curve could be generated to combine size effects and temporal effects on the mechanical behavior of PS nanofibers at room temperature.
Molecular Confinement; Entanglements; Necking; Strain Hardening
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/730862015-01-21T00:00:00ZInteraction between a conical shock wave and a plane compressible turbulent boundary layer at Mach 2.05
http://hdl.handle.net/2142/72992
Interaction between a conical shock wave and a plane compressible turbulent boundary layer at Mach 2.05
The interaction between an impinging conical shock wave with a plane compressible turbulent boundary layer has been studied at Mach 2.05. Surface oil flow and pressure-sensitive paint (PSP) data were obtained beneath the oncoming boundary layer, while schlieren and particle image velocimetry (PIV) data were obtained in the streamwise/wall-normal (x-y) plane. Oil flow data suggested that the interaction causes two-dimensional (2D) separation near the centerline, and outside of this region three-dimensional (3D) separation that propagates fluid away from the centerline toward the sidewall. PSP results showed relatively constant upstream-influence length across the inviscid shock trace. PSP also revealed significant spanwise and streamwise expansion just downstream of the shock trace, unlike the qualitatively similar two-dimensional, wedge-generated oblique shock/boundary-layer interaction. Schlieren data suggested that the flow through the interaction was unseparated, and that there is significant unsteadiness in interaction position away from the centerline due to variation in the incoming boundary layer. PIV data showed the convection of large-scale vortical structures at velocities on the order of the streamwise velocity at the vortex center. These structures were smoothed out in the interaction. The PIV data moreover confirmed the downstream expansion shown by the PSP as well as a mean lack of flow separation. However, PIV suggested that there were some cases of instantaneous separation. Overall, the interaction diverts fluid away from a low pressure zone a little way downstream of the shock. Ultimately, the geometric three-dimensionality of the problem manifests as a preferred three-dimensionality of fluid transport next to the wall, unlike the qualitatively similar, 2D oblique shock-wave/boundary-layer interaction.
Taylor Maccoll; Cone Flow; Shock Boundary Layer Interaction (SBLI); Shock Wave Boundary Layer Interaction; Pressure Sensitive Paint; Particle Image Velocimetry; Supersonic
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/729922015-01-21T00:00:00ZDesign of 3D swept wing hybrid models for icing wind tunnel tests
http://hdl.handle.net/2142/72880
Design of 3D swept wing hybrid models for icing wind tunnel tests
The study of aircraft icing is critical to ensure the safety of any aircraft that might experience icing conditions in flight, including general, commercial, and military aviation. The certification of modern commercial transports requires manufacturers to demonstrate that these aircraft can safely operate during icing conditions through a set of flight tests, consistent with the standards set forth by the Federal Aviation Administration.
This is often expensive and challenging to find the appropriate icing test conditions. Thus, both computational methods and icing wind tunnel experiments are utilized during the design and certification of aircraft ice-protection systems to provide a controlled and repeatable environment to mitigate risks, reduce costs, and validate the existing computational icing tools.
However, the existing icing wind tunnel facilities cannot accommodate large wings such as those found on modern commercial aircraft without being dramatically scaled. Two methods of scaling exist. The first geometrically scales the entire geometry to fit inside the tunnel test section and then scales the icing conditions to obtain icing similitude. The second maintains the full-scale leading edge of the reference geometry and replaces the aft section with a truncated trailing edge that produces a similar flowfield around the leading edge with a significantly shorter chord, reducing model size and tunnel blockage. This type of model is referred to as a hybrid and its biggest advantage lies in the fact that it is designed to produce full-scale ice shapes, while reducing or even eliminating the need for icing scaling. While a design method for a straight, untapered hybrid wing is well documented and there is a broad set of experimental data available, the design of a swept, hybrid wing lacks both a design method and experimental data.
This thesis established a design method for large hybrid swept wings that reproduce full-scale ice accretions through icing wind tunnel tests. The design method was broken down in two steps: 1) A 2D hybrid airfoil design, and 2) A 3D hybrid swept wing design. Multiple existing computational tools were employed and several parametric studies performed.
It was shown, in 2D, that matching the stagnation point location on the leading edge of the hybrid airfoil had a first-order impact on matching the full-scale ice shape, while matching the suction peak magnitude and location had a second-order effect. The closer to the leading edge lift was generated for a given hybrid design, the less total load was required to reach the same stagnation point location. As an implication, more front-loaded airfoils required less lift than more aft-loaded ones to reach the same stagnation point location on a hybrid airfoil. More front load also increased the risk of flow separation near the leading edge, while more aft load increased the risk of separation near the trailing edge. Finally, higher hybrid scale factors were shown to increase the risk of flow separation.
In 3D, sweep angle was shown to be the primary cause for attachment line location spanwise variation, while aspect ratio did not have a significant impact. Matching attachment line location on the leading edge of the hybrid wing model also had a first-order impact on matching ice shape, similar to matching stagnation point location in 2D hybrid airfoils. More front-loaded 3D hybrid wing models not only yielded less total load to reach the same centerline attachment line location, but also showed the additional benefit of reducing the attachment line location spanwise variation. The attempted spanwise load control techniques had different effectiveness on the hybrid wing models. Adding a sidewall gap between the model and the outboard sidewall helped prevent flow separation near the wingtip, but did not effectively change the attachment line slope across span. The use of segmented flaps to equalize load across span was found to be highly dependent on the model aspect ratio and was ineffective for values lower than 2. Additionally, thicker models relatively to the wind tunnel test section yielded more tunnel blockage, presenting a significant effect on suction peak magnitude for values of tunnel height over model chord h/c lower than 2.
Finally, 3 hybrid wing models were designed utilizing the established design method to represent three selected stations of the 65%-scaled Common Research Model, to be tested in the 6 by 9 ft. test section of the NASA Glenn Icing Research Tunnel. 3D aerodynamic and ice accretion simulations were performed utilizing Fluent (3D RANS solver) and LEWICE3D (ice accretion code) to show the successful performance of these models in predicting the full-scale ice accretions for a set of different aerodynamic and icing conditions.
aerodynamics; wing design; aircraft icing; wind tunnel; icing; ice shape; accretion; hybrid; sweep; swept; wing; computational fluid dynamics (CFD); aircraft certification
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728802015-01-21T00:00:00ZModeling of non-equilibrium plasmas in an inductively coupled plasma facility
http://hdl.handle.net/2142/72858
Modeling of non-equilibrium plasmas in an inductively coupled plasma facility
This work presents the results of the numerical simulation of ionized gas ﬂows inside the torch of
an inductively coupled plasma facility in the von Karman Institute for Fluid Dynamics. The main
purpose of this work is the parametric investigation of thermo-chemical non-equilibrium eﬀects on
the plasma jet at diﬀerent operating pressures ranging from 3 to 15 kPa. The test gas is an ionized
air mixture represented by eleven species. The induced electric ﬁeld inside the torch is computed by
solving Helmholtz induction equation. The non-equilibrium eﬀects are modeled using a standard
two-temperature formulation. In particular, the present analysis assesses the impact of diﬀerent
chemical kinetics models and vibration-chemistry-vibration coupling models on the resulting ﬂow
and electro-magnetic ﬁelds. Results of the present work are compared with solutions previously
computed by assuming local thermodynamic equilibrium (LTE)conditions and signiﬁcant diﬀerences
on both the induced electric and temperature ﬁeld are observed especially for the low pressure cases.
Partially ionized gases; Inductively coupled plasma facility; Preferential/non-preferential models; Advection upstream split method
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728582015-01-21T00:00:00ZEffects of micro- and nano-structure on the deformation response of a Ag60Cu40 lamellar and rod-in-matrix eutectic alloy
http://hdl.handle.net/2142/72853
Effects of micro- and nano-structure on the deformation response of a Ag60Cu40 lamellar and rod-in-matrix eutectic alloy
The presented work investigates the mechanical response of the silver-copper eutectic system (Ag60Cu40, subscripts indicating atomic percent) linking material deformation to microstructural properties. The Ag60Cu40 material system can be produced as either a multidirectional lamellar or unidirectional reinforcement-in-matrix micro-structure. Specimens of each micro-structure type were studied under quasi-static and dynamic loading conditions.
The first part of this work focuses on the study of the material with a multidirectional lamellar structure. Materials produced with this structure primarily consist of eutectic colonies of alternating layers of silver and copper with layer thicknesses between 35 nm – 200 nm. The orientations of the eutectic colonies are randomly distributed throughout the material resulting in the formation of boundaries between neighboring eutectic colonies which have different orientations with respect to each other. The strength of this material is shown to be strain rate insensitive over the strain rates studied (10-3 s-1 to 103 s-1). Comparisons are made between the Ag-Cu stress-strain response and literature stress-strain responses of nano-structured silver and nano-structured copper demonstrating the high strength of the multidirectional Ag-Cu system. Three primary deformation mechanisms that occur at increasing levels of strain at the specimen radial surface are identified: kinking, brooming, and interfacial delamination. At the specimen interior kinking is the only mechanism observed.
The second portion of this work examines the unidirectional reinforcement in matrix structure again for Ag60Cu40. This structure has a common crystallographic direction matching the axial direction of the cast material. From a single casting specimens are machined such that loading along three directions oriented 1) parallel to, 2) at 45˚ to and 3) perpendicular to the can occur using dynamic loading. Through alterations in the solidification rate of the unidirectional cast material the micro-structure nominal feature size can be regulated obtaining castings with either 200 nm, 500 nm, 800 nm or 1.2 µm thick reinforcements. For each loading orientation the dynamic material response is presented with the observed internal and external deformation mechanisms. Comparisons of the recorded elastic modulus, yield strength, and strain hardening exponent are made over the loading orientations and nominal micro-structure feature sizes. Crystal anisotropy is used to account for variation in the observed elastic modulus of each loading orientation. Dislocation deformation mechanisms are used to explain the differences in the yield strength and strain hardening. The mechanical properties of the multidirectional lamellar structure are compared to the unidirectional material structure. The multidirectional material is shown to have a higher yield strength. The unidirectional material is shown to have greater strain hardening when microstructure features sizes are greater than 500 nm under certain loading directions.
eutectic system; silver-copper; high-strain rate; deformation mechanisms
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728532015-01-21T00:00:00ZOptimal control problems on lie groups with symmetry breaking cost functions
http://hdl.handle.net/2142/72845
Optimal control problems on lie groups with symmetry breaking cost functions
In this thesis, we consider smooth optimal control systems that evolve on Lie groups. Pontryagin's maximum principle allows us to search for local solutions of the optimal control problem by studying an associated Hamiltonian dynamical system. When the associated Hamiltonian function possess symmetries, we can often study the Hamiltonian system in a vector space whose dimension is lower than the original system. We apply these symmetry reduction techniques to optimal control problems on Lie groups for which the associated Hamiltonian function is left-invariant under the action of a subgroup of the Lie group. Necessary conditions for optimality are derived by applying Lie-Poisson reduction for semidirect products, a previously developed method of symmetry group reduction in the field of geometric mechanics. Our main contribution is a reduced sufficient condition for optimality that relies on the nonexistence of conjugate points. Coordinate formulae are derived for computing conjugate points in the reduced Hamiltonian system, and we relate these conjugate points to local optimality in the original optimal control problem. These optimality conditions are then applied to an example optimal control problem on the Lie group SE(3) that exhibits symmetries with respect to SE(2), a subgroup of SE(3). This optimal control problem can be used to model either a kinematic airplane, i.e. a rigid body moving at a constant speed whose angular velocities can be controlled, or a Kirchhoff elastic rod in a gravitational field.
geometric optimal control; symmetry reduction; conjugate points
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728452015-01-21T00:00:00ZParticle packings and microstructure modeling of energetic materials
http://hdl.handle.net/2142/72841
Particle packings and microstructure modeling of energetic materials
This dissertation explores the use of packings of frictionless hard particles as models of the microstructure of particulate heterogeneous materials.
In the first part of this dissertation, we present the current mathematical framework used for understanding the properties of particle packings,
as well as the methods and algorithms we have developed to generate packings of frictionless hard particles with a computer.
We develop two algorithms to model hard-particle systems: a collision-driven molecular dynamics algorithm for the simulation of packings of spheres,
and a novel hybrid algorithm employing both molecular dynamics and Monte Carlo techniques for the simulation of packings of particles with general convex shapes,
such as spheres, cylinders, ellipsoids, polyhedra, etc.
We focus heavily on performance in order to enable the simulation of large systems containing 10⁶–10⁷ particles, previously too computationally expensive to simulate.
We use performance benchmarks to demonstrate that our implementations of these algorithms scale roughly linearly with the number N of particles in the system,
and show the impact that polydispersivity has on performance.
In the second part of this dissertation we explore the properties of disordered and ordered hard-particle packings.
We reproduce key results found in the literature for packings of spheres and polyhedra, and discuss some of their statistical properties.
We then follow the discussion with applications of particle packings as models of the microstructure of particulate materials obtained via computed tomography.
We find that the shape of the particles and their size distribution both play a crucial role in the determination of the statistical properties of heterogeneous materials.
particle packing; heterogeneous materials; microstructure modeling; energetic materials; solid propellants; polyhedron packing
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728412015-01-21T00:00:00ZExperimental and numerical validation of ion extractor grids
http://hdl.handle.net/2142/72822
Experimental and numerical validation of ion extractor grids
The experimental plasma thruster named the Helicon Injected Inertial Plasma Electrostatic Rocket (HIIPER)
at the University of Illinois at Urbana-Champaign (UIUC) is currently being investigated. HIIPER features
an Inertial Electrostatic Con nement (IEC) grid that is made asymmetrical by cutting out several wires to
enlarge one of the openings in the spherical grid. When fed with a non-fusing gas such as argon, HIIPER
operates in a mode where a very sharp plasma beam forms and exits the system through the IEC's asymmetry,
thus allowing the existence of a net non-zero momentum. Such an operating mode is called "Jet-Mode".
However, it was demonstrated that the exiting plasma beam is mostly dominated by electrons, as opposed to
ions. Since electrons carry a very small momentum due to their light mass, an asymmetrical IEC producing
an electron-dominated beam is not a viable propulsion device. In order to provide a greater thrust when the
IEC is operating in the Jet-Mode, heavier particles such as neutralized ions must exit through the beam.
This thesis focuses on the validation of so-called extractor grids, which are electrostatic components that
can alter the potential pro le inside the system, and therefore enable the escape of heavier particles towards
the preferred exit direction that would otherwise be trapped inside the system. The validation process
mentioned before was two-fold: rst, a 2D, axisymmetric, electrostatic Particle-In-Cell code was developed
in order to benchmark the di erent extractor grid designs and select those which performed the best at
ion extraction. In particular, it was demonstrated that some speci c parameters have a favorable e ect
on particle extraction. Second, an experimental campaign consisting in the testing of the extractor grids
designs, as well as improving the experimental set-up, was carried out. Conclusions were then drawn on the
capacity of HIIPER to act as a space thruster, once augmented with extractor grids. Based on the result of
this work, suggestions for future research are made.
plasma; electric propulsion; Inertial Electrostatic Confinement (IEC); Helicon Injected Inertial Plasma Electrostatic Rocket (HIIPER); extractor grids
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728222015-01-21T00:00:00ZStructure and properties of the interdiffused phase between lithiated silicon and copper current collector
http://hdl.handle.net/2142/72820
Structure and properties of the interdiffused phase between lithiated silicon and copper current collector
Silicon is an important electrode material for next generation high performance lithium ion batteries due to its order of magnitude higher charge carrying capacity compared to conventional graphite electrodes. The main obstacle of using Si electrodes in commercial lithium batteries is the massive volume expansion of the Si electrode under repeated charge cycling, which leads to delamination of the Si electrode from the Cu current collector and inevitably results in capacity fade. Using first principle calculations based on density functional theory and ab-initio molecular dynamics simulations, this thesis focuses on the structure of the Cu/Si interface and aims to provide a complete picture of the intermixing at the Cu/Si interface during lithiation processes. The hypothesis, supported by existing experiments, is that the Cu/Si interface is not pristine and comprises of an interdiffused Li-Si-Cu interphase structure. To test this hypothesis, the barrier energies for Li diffusion into the assumed crystalline silicide interphase structure separating crystalline Si and Cu is studied. Results show that the barrier energies for Li ion diffusion decreases towards the interphase structure, which suggest that Li ions can diffuse into the silicide structure even during early stages of lithiation. Several interdiffused Li-Si-Cu interphase structures with varying Li to Si content are subsequently modeled using rapid heating and quenching process. The atomic structure of interdiffused Li-Si-Cu phase reconstructed from rapid heating and quenching are in good agreement with previous experiment results. The work of separation of these interdiffused phase are also examined.
Lithium battery; interdiffused; Si/Cu interface
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/728202015-01-21T00:00:00ZWave tailoring granular materials: effect of randomness and plane wave propagation
http://hdl.handle.net/2142/72754
Wave tailoring granular materials: effect of randomness and plane wave propagation
This thesis addresses some of the fundamental issues/aspects as well as practical significance of impact response in granular media. In the first part of the study, we investigate numerically the effect of randomness in material and geometric properties on wave propagation behavior in 1D and 2D granular media. Results obtained for both 1D chains and 2D media show the kinetic energy amplitude decays with distance, with the rate of decay found to depend on the level of randomness and the distance from point of impact. The kinetic energy amplitude initially decays exponentially before transitioning to a universal power-law regime that is valid for all levels of randomness. The power-law regime is fundamentally due to the presence of secondary waves whose amplitude is higher than that of the primary wave after the point of transition. Another key result quantifies the rate of decay of force amplitude in a 2D square packing system along various directions of propagation. Several contour maps are obtained that demonstrate the directions along which we can obtain minimum or maximum decay, which are practically relevant for material design.
In the second part of the study, we focus on plane wave propagation in higher dimensional structures. In the case of 2D and 3D monodisperse granular media, we demonstrate an equivalence with 1D chains and consequently derive the relation between wavefront speed and force amplitude in higher dimensional systems. Subsequent normalization results in a universal wavefront speed-force amplitude relation that is valid across the different ordered 2D and 3D systems such as hexagonal, body-centered cubic and face-centered cubic packings. We also investigate the effect of angular impact on granular media and discuss mechanism through which the shearing component of the loading is propagated in the system. In the case of 2D dimer systems, we consider a square packing system with interstitial intruders. Following the procedure that we developed for monodisperse granular media, we obtain an equivalent nonlocal dimer chain that gives the same response with relevant scaling of material properties. In this study, we demonstrate the existence of a new family of plane solitary waves over a wide range of material and geometric properties. We also indicate a discrete set of solutions for which there is locally maximum decay, thereby showing promise for wave mitigation as well.
In the last part of our study, we conduct a preliminary study to investigate the vibration response of beams made of a granular chain embedded in a linear elastic matrix. A nonlinear, dynamic finite element model is developed, in which the granular chain is converted to a series of 1D nonlinear bar elements whose contribution is added to linear quadrilateral elements. We study the bending response by applying a harmonic loading on a composite beam fixed on both ends and show a fundamental difference between the dynamic response of a linear elastic beam and the embedded granular system at resonance. Unlike a linear elastic beam, we find the deflection of an embedded granular system to be finite at resonance. Furthermore, in the presence of precompression, the frequency at which the composite beam deflection is maximum can be controlled based on the level of precompression, acting as an active control feature in material design.
Granular materials; wave propagation; solitary wave; randomness; vibrations
Wed, 21 Jan 2015 00:00:00 GMThttp://hdl.handle.net/2142/727542015-01-21T00:00:00Z