book of abstracts...book of abstracts sibley graduate research symposium show us why your research...

Book of Abstracts

Sibley Graduate Research

Symposium

Show us why your research is awesome!

January 20, 2017

The Sibley School of

Mechanical & Aerospace Engineering

Cornell UniversityIthaca, NY

About

The Sibley Graduate Research Symposium (SGRS) is astudent run symposium that strives to promote the sharingof work among di�erent areas of research to foster academiccollaboration. SGRS provides a platform for students tointroduce and be introduced to exciting research, as well asbenefit from conference participation and attendance, allwhile promoting the Sibley Graduate School.

SGRS Committee Members: Robert Chiodi, Joseph Corbett-

Davies, Stephanie Firehammer, Alex Ivanov, Reece Kearney,

Adam Pacheck, Kate Polhemus, Nate Rogers, Ryan Snodgrass,

Gabriel Soto, Sasank Vemulapati, Patricia Xu

SGRS Advisory Board: Prof. Brian Kirby, Dr. Marcia Sawyer

©Sibley Graduate Research Symposium 2017, Sibley School ofMechanical & Aerospace Engineering, Cornell University, Ithaca,NY.

ii

Contents

Contents iii

Schedule 2

Keynote Speaker 9

Presentation Abstracts 13

1 Engineering for Health Diagnostics 131.1 A portable multi-colored micronutrient di-

agnostic platform for iron, vitamin A andacute inflammation status assessment . . . 14

1.2 Acoustic waves-based microfluidic device to-wards separation of vesicle subpopulationsderived from cancer cells . . . . . . . . . . . 15

1.3 Portable health diagnostics for the develop-ing world . . . . . . . . . . . . . . . . . . . 16

1.4 Stem cell di�erentiation and teratoma-formingpotential in hiPSC-derived neural cultures . 17

2 Computational Fluid Dynamics 192.1 Computational Exploration of Air-Blast At-

omization . . . . . . . . . . . . . . . . . . . 20

iii

2.2 A pre-partitioned adaptive chemistry method-ology for e�cient implementation of combus-tion chemistry in LES/PDF methods . . . . 21

2.3 Rheology & hydrodynamic di�usion in sus-pensions of aligning ring-shaped particles ina low Reynolds number simple shear flow . 22

3 Optical Systems and Signal Processing 233.1 Misalignment Retrieval of an O�-axis Parabolic

Mirror using Batch Estimator and KalmanFilter . . . . . . . . . . . . . . . . . . . . . 24

3.2 Blind Source Separation Algorithms for StarPSF Subtraction in Planetary Direct Imaging 25

3.3 Modulation Instability in Multimode OpticalFiber . . . . . . . . . . . . . . . . . . . . . . 26

4 Biomechanics 274.1 Generating Mechanical Stresses in Bacteria

with a Microfluidic Platform . . . . . . . . . 284.2 Raman and FTIR bone quality parameters

correlate with physical chemical propertiesof model compounds and native bone tissue 29

4.3 The Mechanism of Silica Nanoparticle PE-Gylation Processes . . . . . . . . . . . . . . 30

5 Experimental Fluids 315.1 Manipulating Flow Through Wind Farms To

Increase Their E�ciency: An ExperimentalInvestigation . . . . . . . . . . . . . . . . . 32

5.2 Impingement of a Counter-Rotating VortexPair on a Wavy Wall . . . . . . . . . . . . . 33

5.3 Improving Oil Detection and Extraction Ef-ficiency by Using Hairy Nanoparticles . . . 34

6 Dynamic Systems and Robotics 356.1 The Physics of Motorcycles and Fast Bicy-

cles: Lean, Stability and Counter-steering . 36

iv

6.2 A Blade Element Approach to ModelingAerodynamic Flight of an Insect-scale Robot 37

6.3 Designing and Optimizing Satellite Constel-lations: Applying and Extracting LessonsLearned . . . . . . . . . . . . . . . . . . . . 38

6.4 A Hierarchical Framework for GuaranteedPerformance in the Multi-Robot Multi-TargetTracking Problem . . . . . . . . . . . . . . . 39

6.5 Starshade Orbital Maneuver Study for WFIRST 40

Sponsors 43

v

Schedule

1

Room: Upson 431

Engineering for Health Diagnostics [12:00 - 1:00 PM]

12:00 PM A portable multi-colored micronutrientdiagnostic platform for iron, vitaminA and acute inflammation status as-sessment

12:15 PM Acoustic waves-based microfluidic de-vice towards separation of vesicle sub-populations derived from cancer cells

12:30 PM Portable health diagnostics for the de-veloping world

12:45 PM Stem cell di�erentiation and teratoma-forming potential in hiPSC-derivedneural cultures

BREAK [1:00 - 2:00 PM]

Optical Systems and Signal Processing[2:00 - 2:45 PM]

2:00 PM Misalignment Retrieval of an O�-axisParabolic Mirror using Batch Estima-tor and Kalman Filter

2:15 PM Blind Source Separation Algorithmsfor Star PSF Subtraction in PlanetaryDirect Imaging

2:30 PM Modulation Instability in MultimodeOptical Fiber

2

Room: Upson 531

Computational Fluid Dynamics [12:00 - 12:45 PM]

12:00 PM Computational Exploration of Air-Blast Atomization

12:15 PM A pre-partitioned adaptive chemistrymethodology for e�cient implemen-tation of combustion chemistry inLES/PDF methods

12:30 PM Rheology & hydrodynamic di�usionin suspensions of aligning ring-shapedparticles in a low Reynolds numbersimple shear flow

BREAK [1:00 - 2:00 PM]

Biomechanics [2:00 - 2:45 PM]

2:00 PM Generating Mechanical Stresses in Bac-teria with a Microfluidic Platform

2:15 PM Raman and FTIR bone quality param-eters correlate with physical chemicalproperties of model compounds andnative bone tissue

2:30 PM The Mechanism of Silica NanoparticlePEGylation Processes

3

Room: Upson 431

Experimental Fluids [3:00 - 3:45 PM]

3:00 PM Can changing the set point of a windturbine accelerate the decay of its mo-mentum deficit?

3:15 PM Impingement of a Counter-RotatingVortex Pair on a Wavy Wall.

3:30 PM Improving Oil Detection and Ex-traction E�ciency by Using HairyNanoparticles

4

Room: Upson 531

Dynamic Systems and Robotics [3:00 - 4:15 PM]

3:00 PM The Physics of Motorcycles and FastBicycles : Lean, Stability and Counter-steering

3:15 PM A Blade Element Approach to Model-ing Aerodynamic Flight of an Insect-scale Robot

3:30 PM Designing and Optimizing SatelliteConstellations: Applying and Extract-ing Lessons Learned

3:45 PM A Hierarchical Framework for Guaran-teed Performance in the Multi-RobotMulti-Target Tracking Problem

4:00 PM Starshade Orbital Maneuver Study forWFIRST

5

Keynote Speaker

7

Jacque Cole, Ph.D.

The Therapeutic Importance of Vasculature onBone Health

Thurston Room No: 205, Time: 10:00 am

Abstract: Vascular function is vital for bone health,yet it is often overlooked in skeletal research. Bone willnot develop or heal without a functional vascular net-work, and a number of clinical conditions involve bothvascular decline and bone loss, including aging, obe-sity, and stroke. However, exactly how the vasculatureinteracts with bone is not well understood, such as howparticular aspects of bone are altered by changes invascular signaling and blood flow. My lab perturbs vas-culature systemically with established mouse models ofobesity and stroke to determine the impact of vascularchanges on bone metabolism and mechanics. We be-lieve that osteovasculature is a novel therapeutic targetfor treating and preventing bone loss in several clinicalcontexts. Understanding more about the expressionof various cytokines, particularly those involved intissue crosstalk (e.g., bone-vasculature, muscle-bone,muscle-vasculature), will expand our understanding ofthe mechanisms involved in these processes and willbe critical for developing drugs or physical therapiesto mitigate bone loss with vascular-related conditions

9

like obesity and stroke. Understanding how osteovas-culature changes with obesity and stroke and how thatimpacts bone metabolism and mechanics, as well asthe physiological mechanisms involved with exerciseor exercise-like interventions, are critical first steps fordeveloping new rehabilitation strategies that addressthe bone loss experienced with these conditions.

Biosketch: Jacque Cole is an Assistant Professor inthe Joint Department of Biomedical Engineering atNorth Carolina State University and the Universityof North Carolina-Chapel Hill, and she is the direc-tor of the Orthopaedic Mechanobiology Lab (OML).She received a Bachelor of Mechanical Engineering(B.M.E.) degree from Auburn University in 2001 andan M.S. and Ph.D. in Mechanical Engineering (withminors in Biomedical Engineering and Biometry) fromCornell University in 2004 and 2007, respectively. Shethen completed two postdoctoral fellowships at theUniversity of Michigan in Chemistry (2007-2010) andOrthopaedic Surgery (2010-2013). In 2015, she re-ceived an NIH K12 award through the NorthwesternUniversity Program in Movement and RehabilitationSciences to study the physiological mechanisms ofbone loss following stroke, and she recently receivedan NIH R21 to investigate the parallel development ofbone and muscle impairments in the shoulder follow-ing neonatal nerve injury, a collaborative project withDr. Kate Saul (NC State Mechanical & AerospaceEngineering). The overall goal of Dr. Cole’s researchis to improve current therapeutic strategies for treat-ing and preventing bone loss associated with aging,disease, and injury.

10

Presentation Abstracts

11

Section 1

Engineering for HealthDiagnostics

13

A portable multi-colored micronutrientdiagnostic platform for iron, vitamin A and

acute inflammation status assessment

Zhengda Lu

Advisor: Prof. David Erickson

Upson Room No: 431, Time: 12:00 pm

Vitamin A and iron deficiency are two of the most preva-lent micronutrient deficiencies in the world. VitaminA deficiency causes night blindness, impaired immunefunction, and birth defects, while iron deficiency leadsto anemia, reduced physical work capacity, and impairedbrain development. Vitamin A and iron deficiency of-ten coexist, particularly in the non-industrialized world,and can exacerbate each other. In this work, we devel-oped a multiplexed, point-of-care, low cost diagnosticplatform capable of simultaneously diagnosing vitamin Aand iron deficiencies in about 10 minutes. The platformcomprises a multi-colored fluorescence lateral flow assaytest strip, and a portable reader that be controlled by anymobile device with WiFi connectivity. In the platform,retinol binding protein and ferritin are used as surro-gate biomarkers for vitamin A and iron deficiency, andC-reactive protein is included to exclude invalidate resultscaused by inflammation. We demonstrated the e�cacy ofthe platform by accurately quantifying concentrations ofthe three biomarkers simultaneously in standard solutions.Furthermore, we developed an innovative fluorescence sig-nal discriminating algorithm and proved its capabilityto quantitatively measure levels of three biomarkers in asingle detection area.

14

Acoustic waves-based microfluidic devicetowards separation of vesicle subpopulations

derived from cancer cells

Prateek Sehgal

Advisor: Prof. Brian Kirby


Extracellular shed vesicles (ESVs) are a distinct intercellu-lar signaling modality with ubiquitous function which, in acancer context, have been shown to promote tumorigenesis.Critically, cancer-cell-derived ESVs exhibit two relevantsubpopulations: microvesicles (MVs) ranging from 200nm-1um and exosomes ranging from 30nm-100nm. Our grouphas shown that larger vesicles are cancer-specific and tar-getable with metabolic inhibitors, however, the di�erencesin cargo between these subpopulations are not known andthus the paths to therapeutic manipulation of these vesi-cles are currently fuzzy. Although techniques for vesicleisolation exist (e.g. centrifugation and filtration, often inconjunction with immunospecific agglomeration or tag-ging), these approaches fail owing to the fragility anddensity variations of the vesicles. In this work, we presenta surface acoustic wave-based microfluidic device in anovel geometry to isolate undamaged functionally-intactmicrovesicles from ESV-enriched solutions based on vesiclesize. We present a theoretical framework to stochasticallymodel nanoparticle migration in acoustophoresis-baseddevice and analytically demonstrate the nanoparticle sep-aration modality of our device. We also present the com-putational analysis of acoustic fields used to predict thefunctionality of the device. Lastly, we experimentallydemonstrate the working principle of our device usingfluorescent polystyrene nanoparticles to show its potentialto isolate nanosize microvesicles from exosomes withoutdamaging the vesicles.

15

Portable health diagnostics for thedeveloping world

Ryan Snodgrass

Advisor: Prof. David Erickson


Nucleic acid based assays are a very sensitive and specifictool for disease diagnosis, but are usually constrained tothe laboratory setting because of the need for expensiveequipment and reliable electricity. Over the past years, ourteam has been at work building a device that is capableof performing these nucleic acid based assays in the fieldwithout any laboratory equipment. We are motivatedin particular by a HIV-derived cancer called Kaposi’ssarcoma, most common today in sub Saharan Africa. InJuly, 2016, we took a prototype of our portable diagnosticdevice to Uganda and conducted preliminary testing. Inthis talk, I will discuss our visits to healthcare clinics inrural Uganda, the design and operation of our diagnostictool, and the fundamentals of the nucleic acid assay. Thedevice is heated solar thermally, and is maintained at thecorrect temperature for nucleic acid amplification usingphase change materials. We use simple, cost e�ectiveoptical and electrical parts to achieve a measurement ofviral nucleic acid content, allowing us to both diagnose apatient and interpret disease progression.

16

Stem cell di�erentiation andteratoma-forming potential in hiPSC-derived

neural cultures

Kyle Wellmerling



Currently up to 1 billion people worldwide su�er from dis-abling neurodegenerative disorders and traumatic nerve in-juries. Patient-specific, human-induced pluripotent stemcells (hiPSCs) represent a highly promising strategy toderive di�erentiated autologous neural cells. However,non-specific contaminant cells limit the therapeutic trans-plant potential of neural cells. Current cell sorting meth-ods such as magnetic beads or fluorescence-activated cellsorting lack the specificity required to create contaminant-cell free populations. Therefore, there is an urgent unmetneed for a simple, high-precision technology that can e�-ciently deplete contaminant cells. By developing a simpleintegrated technology to first selectively detach neuralcells via fluid shear stresses, and then employing geomet-rically enhanced di�erential immunocapture to captureresidual contaminant cells, neurotherapeutic translationof hiPSC-derived neural cells will be enabled.

17

Section 2

Computational FluidDynamics

19

Computational Exploration of Air-BlastAtomization

Robert Chiodi

Advisor: Prof. Olivier Desjardins


The air-blast atomization of a planar liquid layer is acomplex fluid phenomenon involving the destabilization ofa low speed liquid layer by a high speed gas coflow. Whileprogress has been made in recent years on understandingthe instability of the liquid surface, it remains di�cultto accurately predict using stability analysis and requiresspecial expertise and equipment to perform thoroughexperiments. Simulations provide an excellent way toconduct parametric studies to determine the e�ect ofsplitter plate geometry and momentum flux ratio on thefrequency and wavelengths of instability, however, areextremely di�cult due to the high density ratio and largerange of length and time scales present in the flow. Usingan accurate conservative level set method in conjunctionwith a newly reformulated reinitialization equation, weperform 3D simulations of the air-blast atomization ofa planar liquid layer and compare them to experiments.We then go on to explore the role momentum flux ratioplays in the longitudinal and transverse wavelengths ofinstability.

20

A pre-partitioned adaptive chemistrymethodology for e�cient implementation ofcombustion chemistry in LES/PDF methods

Ashish Newale

Advisor: Prof. Perrine Pepiot


Large eddy simulation/probability density function (LES/PDF)methods have been shown to be extremely e�ective at ac-curately simulating turbulent flames with strong turbulence-chemistry interaction. A key drawback of particle LES/PDFmethods are their high computational cost comparedto other methods such as flamelet progress variable ap-proach. To significantly reduce the computational expenseof LES/PDF, we propose a pre-partitioned adaptive chem-istry (PPAC-DRGEP) methodology. This methodologyincludes an o�ine preprocessing step in which reducedmodels are generated using directed relation graph witherror propagation (DRGEP) for a preset number of par-titions of the accessed region of the high dimensionalcomposition space. At runtime, the reaction fractionalstep for each particle is performed using an appropriatereduced model, that is identified based on its composi-tion. A LES/PDF code incorporating this PPAC-DRGEPmethodology is used to simulate the experimentally wellcharacterized Sandia flame D. The performance of thismethodology is studied in terms of usage of the reducedmodels and errors incurred during the simulation. Thecomputational savings in terms of runtime are quantifiedto illustrate the potential of the combined PPAC-DRGEPLES/PDF approach.

21

Rheology & hydrodynamic di�usion insuspensions of aligning ring-shaped particles in

a low Reynolds number simple shear flow

Neeraj Sinai Borker

Advisor: Prof. Donald Koch


A class of finite aspect ratio rigid particles that align inthe flow-vorticity plane in an unbounded simple shearflow at low Reynolds number were recently discovered.These flow-aligning particles have the topology of a ringof radius R with a non-circular cross-section which has asharp outer edge and a blunt inner edge. The rheology andhydrodynamic di�usivity in a dilute (nR3«1) suspension offlow aligning rings is obtained from an analysis of pairwisehydrodynamic interactions. These interactions betweenparticles are obtained using a slender-body theory (SBT)developed for the case where the ring radius R is largecompared with the characteristic radius (a) of the cross-section. A method of reflections expansion that can betruncated based on the weak hydrodynamic disturbancesof slender rings along with force and torque free conditionslead to analytical expressions for the ring velocities androtation rates. The asymmetric shape of the ring cross-section leads to a force per unit circumference causedby the local shear flow, a term that does not arise inprevious applications of SBT. It is found that pairwiseinteractions do not cause flow-aligning particles to undergoa full flipping motion through a Je�ery half-orbit. Theseresults show that a modification of particle shape canyield materials with strongly anisotropic structure evenin the absence of external fields, suggesting a route toproducing highly anisotropic reinforced polymers usingsimple processing flows such as injection molding.

22

Section 3

Optical Systems and SignalProcessing

23

Misalignment Retrieval of an O�-axisParabolic Mirror using Batch Estimator and

Kalman Filter

Joyce Fang

Advisor: Prof. Dmitry Savransky


Automation of alignment tasks can provide improvede�ciency and greatly increase the flexibility of an op-tical system. Current optical systems with automatedalignment capabilities are typically designed to includea dedicated wavefront sensor. We demonstrate a self-aligning method for a reflective system using only focalplane images. We define an optical system with a mis-aligned o�-axis parabolic mirror. We perturb the mirrorto generate phase diversity, which allows us to retrievethe misalignments from focal plane images. We performa principal component analysis (PCA) on the simulateddataset to obtain Karhunen-Loeve (KL) modes, whichform the basis set whose weights are the system mea-surements. A model function which maps the state tothe measurement is learned using nonlinear least squaresfitting and serves as the measurement function for thenonlinear estimator. We compare the state estimationresult of a batch estimator and a Kalman filter.

24

Blind Source Separation Algorithms for StarPSF Subtraction in Planetary Direct Imaging

Jacob Shapiro


Upson Room No: 431, Time: 2:15 pmThe principal di�culty with detecting planets via directimaging is that the target signal is similar in magnitude,or fainter, than the noise sources in the image. To compen-sate for this, several methods exist to subtract the PSF ofthe host star and other confounding noise sources. Oneof the most e�ective methods is Karhunen-Loeve ImageProcessing (KLIP). The core algorithm within KLIP isPrincipal Component Analysis, which is a member of aclass of algorithms called Blind Source Separation (BSS).We examine three other BSS algorithms that may po-tentially also be used for PSF subtraction: IndependentComponent Analysis, Stationary Subspace Analysis, andCommon Spatial Pattern Filtering. The underlying prin-ciples of each of the algorithms is discussed, as well asthe processing steps needed to achieve PSF subtraction.The algorithms are examined both as primary PSF sub-traction techniques, as well as additional postprocessingsteps used with KLIP.These algorithms have been used on data from the GeminiPlanet Imager, analyzing images of — Pic b. To build areference library, both Angular Di�erential Imaging andSpectral Di�erential Imaging were used. To compare toKLIP, three major metrics are examined: computationtime, signal-to-noise ratio, and astrometric and photo-metric biases in di�erent image regimes (e.g., speckle-dominated compared to Poisson-noise dominated). Pre-liminary results indicate that these BSS algorithms im-prove performance when used as an enhancement forKLIP, and that they can achieve similar SNR when usedas the primary method of PSF subtraction.

25

Modulation Instability in Multimode OpticalFiber

Zimu Zhu

Advisor: Prof. Frank Wise


The concept of nonlinear wave propagation underlies allsorts of physical systems, from electrodynamics and hy-drodynamics to neuroscience and more. These systemscan all be described by nonlinear partial di�erential equa-tions. Intrinsic to many of them is an instability thatleads to modulation of the steady state, called modula-tion instability. Modulation instability is a remarkablyuniversal phenomenon, which is responsible for all sortsof pattern formation processes.In optical fibers, modulation instability appears as thebreakup of a continuous wave into a train of short pulses.These form solitons, waves that maintain their shapeand width as they propagate. Solitons are remarkableas well they are stable global nonlinear attractors thatspontaneously appear in many systems, such as waterwaves and the collective motion of proteins and DNA.Nonlinear dynamics have been studied extensively in thecontext of single-mode optical fiber (ones that supportonly one guided spatial eigenmode: e�ectively an on-axisbeam). With increasing interest in multimode fiber asmeans of achieving higher power fiber lasers, telecom-munication systems with higher data capacity, and newimaging techniques, there is a need for basic fundamentalstudies. In this work, we explore modulation instabilityin multimode fiber, how it relates to multimode solitons,and its implications towards potential applications of mul-timode fiber.

26

Section 4

Biomechanics

27

Generating Mechanical Stresses in Bacteriawith a Microfluidic Platform

Melanie Roberts

Advisor: Prof. Chris Hernandez


Bacteria must sense and respond to experienced mechani-cal stresses to survive harsh environments. Not much isknown about bacterial response to mechanical stimuli. Ex-isting experimental techniques to investigate mechanosen-sitive pathways require an alteration to the cell physiology,immobilization, or cannot examine the cell on a subcellu-lar level. Here, we demonstrate a method of generatingmechanical stress in live bacteria. Bacteria cells are loadedinto a microfluidic device and trapped in tapered channelswith inlets wide enough to permit entrance but exits onlya fourth of the cell diameter, prohibiting exit (250 nmfeature size). Around the tapered channel is a bypasschannel that experiences a pressure drop, creating a gra-dient in mechanical stress along the length of the cell wall.The flow rate of the liquid medium is adjustable, allowingcontrol of the pressure drop experienced across the cell.The resulting deformation of the bacteria cell is observedduring loading by either traditional or super-resolutionmicroscopy. Once imaged, the measured deformation isput into an analytical mechanical model derived from alinear elastic constitutive equation. The analysis showsthat bacteria under extrusion loading experience predom-inately tensile stresses which increase linearly, where thestress on the downstream end of the cell is on the order ofone megapascal larger than the upstream end. With thisdevice, subcellular characterization of mechanical stressin rod-like bacteria is possible.

28

Raman and FTIR bone quality parameterscorrelate with physical chemical properties of

model compounds and native bone tissue

Erik Taylor

Advisor: Prof. Eve Donnelly

Upson Room No: 531, Time: 2:15 pmRaman spectroscopy is a promising technique for non-invasive characterization of bone composition because ofits ability to characterize hydrated tissues. Thus, thismodality has the potential to address the current needfor in vivo techniques to characterize bone tissue compo-sitional properties beyond the current clinical standardmodality for fracture risk prediction: dual energy x-ray ab-sorptiometry. In particular, local tissue mineral content,crystal maturity, and extent of carbonate substitutionare critical compositional parameters that reflect the ef-fects of disease, age, and drug treatment, and influencestructural, material, and mechanical properties. However,the Raman measures of these parameters have not yetbeen validated. In this study, our objective was to val-idate the Raman mineral:matrix ratio, crystal maturitymeasure, and carbonate:phosphate ratio by correlatingthem to analytical measures in chemical standards andnative bone tissue. Two series of chemical standards con-sisting of 1) varying ratios of synthetic hydroxyapatiteand collagen and 2) varying ratios of carbonate substi-tuted synthetic hydroxyapatite, as well as bone tissuefrom humans, sheep, and mice were characterized withRaman and FTIR spectroscopy, gravimetric analysis, x-ray di�raction, and elemental analysis. Initial resultsshow strong correlations between Raman mineral:matrixand carbonate:phosphate ratios and analytical measures.Together these results connect physical meaning to theseoutcome parameters, an important step in the evolutionof Raman spectroscopy as a technique with potential fornon-invasive characterization of bone composition.

29

The Mechanism of Silica NanoparticlePEGylation Processes

Duhan Zhang

Advisor: Prof. Ulrich Wiesner


Surface modification with polyethylene glycol (PEG; PE-Gylation) is a widely used technique to improve nanopar-ticle (NP) stability, biocompatibility, and biodistributionprofiles. In particular, PEGylation of silica surfaces andcoatings plays a pivotal role across various classes of NPs.Despite the use of numerous protocols there is limitedfundamental understanding of the mechanisms of theseprocesses for NPs. Here, after reaction optimization forparticle stability, we employ several characterizations onultrasmall (10 nm) fluorescent silica nanoparticles (SNPs)in water as a test bed. We show unexpected fast reac-tion kinetics in successful PEGylation observed even atnanomolar concentrations and attributed this to instantnoncovalent adsorption of PEG molecules to the SNPsurface preceding covalent attachment. Further studies ofvarious reaction conditions enable the elucidation of pro-cess design criteria for NP PEGylation and surface modi-fication with functional ligands, which may be applicableto a broad range of NPs thereby accelerating progress infields ranging from biosensing to nanomedicine.

30

Section 5

Experimental Fluids

31

Manipulating Flow Through Wind Farms ToIncrease Their E�ciency: An Experimental

Investigation

Dan Houck

Advisor: Prof. Edwin Todd Cown


Wind farms require large footprints in order to spaceturbines widely enough to mitigate their wakes, whichreduce the power output of downstream turbines. In-terestingly, though, power output does not continue todrop but plateaus downstream due to vertical entrain-ment of energy, which replenishes the supply of energy todownstream turbines. This energy gradient along withanalogous flows such as canopy flows and flow throughbeds of bivalves informs two planned experiments using asmall scale model wind farm in the wide flume.

32

Impingement of a Counter-Rotating VortexPair on a Wavy Wall

Sarah Morris

Advisor: Prof. C.H.K. Williamson


In this research we examine the impingement of a vortexpair onto a wavy wall. Isolated vortex pairs, not in grounde�ect, can become unstable to short-wave (Widnall, 1974)or long-wave instability (Crow, 1970). When a vortex pairapproaches a ground plane, the boundary layer that formson the surface separates, generating secondary vorticityand causing the primary pair to rebound. When a vortexpair with the long-wave instability interacts with a flatboundary, the topology of the pair changes, resulting inrebounding vortical structures whose form is dependenton the extent of the instability prior to wall interaction(Asselin & amp; Williamson, 2016). By using PIV andLIF to consider the "complementary" experiment of astraight vortex pair encountering a wavy wall (ratherthan a wavy pair impinging on a flat wall), certain criticalfeatures of the two flows are found to be similar. The 2Dvortex pair first interacts with the hills of the boundary,triggering accelerated vorticity cancellation in this areacompared to the corresponding valley regions. An axialpressure gradient forms between the two regions, givingrise to strong axial flow. This leads to the interaction ofprimary and secondary vortices in the valleys, whereinreconnection results in "rebounding" vortex rings, two perfundamental wavelength. The resulting flowfield formsdistinctly di�erent vortex structures than are classicallyfound for 2D vortex pair wall impingement or for thelong-wave instability out of ground e�ect.

33

Improving Oil Detection and ExtractionE�ciency by Using Hairy Nanoparticles

Katherine Polhemus



With the rapid depletion of known oil reserves, detectingproperties of the oil reservoirs and optimizing oil extrac-tion is critical. By measuring the aqueous properties ofthe reservoirs, decisions can be made on which reservoirsto drill and the available quantity of oil to extract, withminimal environmental impact. Utilizing hairy nanoparti-cles in testing can provide a variety of information aboutthe reservoir. The objective of the proposed work is tocharacterize the behavior of hairy nanoparticles at theoil-water interface in order to optimize their use as sub-surface sensors. To begin the characterization, the spatialdistribution of the hairy nanoparticles in relation to theinterface needs to be identified. To study characteriza-tion a straightforward and controlled model needs to bedeveloped in order to take precise measurements.

34

Section 6

Dynamic Systems andRobotics

35

The Physics of Motorcycles and Fast Bicycles:Lean, Stability and Counter-steering

Shayak Bhattacharjee

Advisor: Prof. Richard Rand


In this lecture I will present the nonlinear equation ofmotion (EOM) of a motorcycle or fast bicycle for the firsttime in Literature. I will use this equation to calculatethe lean angle during a turn, prove that the straight orturning motion of the vehicle is stable, and present amathematical model of counter-steering, the phenomenonby which the vehicle entering or exiting a turn first brieflygoes the wrong way and then starts going the right way.Although a vast literature exists on this topic, it containsonly linear equations, obtained from ad hoc linearizationabout a straight or turning state. These equations allpossess an instability in the lean angle or capsize model,which is not observed in reality. The technical di�culty ofthe analysis arises from the inclusion of the heavily non-holonomic constraints in the Lagrangian equations. Here Iwill instead use a direct force and torque balance approachto the problem to obtain a third order nonlinear ODE forthe vehicle. I will prove that the eigenvalue correspondingto the capsize mode is in fact exactly zero, and robustlyso. I will also obtain quantitatively the transitions intoand out of turns, which clearly show counter-steeringbehaviour.

36

A Blade Element Approach to ModelingAerodynamic Flight of an Insect-scale Robot

Taylor Clawson

Advisor: Prof. Silvia Ferrari


The RoboBee is an insect-scale flying robot that takes in-spiration from insect flapping flight and can be fabricatedusing the smart composite microstructures methodologydeveloped by the Harvard Microrobotics Lab. Most dy-namic models to date have assumed stroke-averaged forcesdue to the negligible body rotation caused by individualwing strokes. This paper incorporates blade element the-ory with a rigid body dynamics model to predict instan-taneous aerodynamic forces during each wing stroke andbetter account for changes in aerodynamic force duringflight.

37

Designing and Optimizing SatelliteConstellations: Applying and Extracting

Lessons Learned

Nozomi Hitomi

Advisor: Prof. Daniel Selva


Decision support tools are beginning to play an activerole in the design of complex space systems such as dis-tributed spacecraft missions, which can simplify eachspacecraft, reduce launch costs by creating spacecraftthat are secondary payloads, increase robustness to to-tal mission failure, achieve global coverage, and reducerevisit times. Designing DSMs, however, are di�cult todesign, especially in the case of heterogenous DSMs, be-cause of the 1) large number of design variables and theircomplex, nonlinear interactions, 2) many nonlinear con-straints involving packaging space, power requirements,link budgets, thermal control, and on-board data handlingcapabilities, 3) multiple conflicting objectives of maximiz-ing performance while minimizing cost and risk, and 4)local optima present in the tradespace. Multiobjectiveevolutionary algorithms (MOEA) are typically used tosolve such complicated real-world problems, but they areconsidered computationally ine�cient because it does notutilize information about the problem structure or the do-main. This work presents a framework that significantlyimproves the e�ciency of MOEAs by incorporating expertknowledge into the algorithm and employing data miningto extract new knowledge that can guide the algorithmtoward the promising regions of the tradespace. Theframework will be demonstrated on a design problem fora cost-e�ective climate-monitoring satellite constellation.

38

A Hierarchical Framework for GuaranteedPerformance in the Multi-Robot Multi-Target

Tracking Problem

Alex Ivanov

Advisor: Prof. Mark Campbell


Muti-agent multi-target tracking is a uniquely di�cultproblem for autonomous robots. Robotic exploration andtracking have traditionally been considered as two sep-arate, but related problems which are usually solved intwo phases: exploration then tracking. The explorationproblem is usually viewed through an information theo-retic framework where a robotic agent attempts to gatheras much information about the environment or an Objectof Interest (OI). Conversely, the target tracking problemattempts to maintain precise location information aboutan OI over time. In any real scenario, these two tasksare sequentially dependent. In other words, autonomousagents must first explore a space to locate OIs and mustsubsequently track these objects. This work proposesa framework which enables the multi-robot multi-targetproblem to be solved in a single phase. The frameworkallows organic autonomous development from the explo-ration problem to the tracking problem as more OIs arediscovered. The hierarchical nature of this frameworkalso enables the coordination of multiple robotic agents intracking multiple OIs to remain computationally e�cient.In addition, probabilistic guarantees are utilized to ensuretracking performance throughout the exploration/trackingtask.

39

Starshade Orbital Maneuver Study forWFIRST

Gabriel Soto



The Wide Field Infrared Survey Telescope (WFIRST) mis-sion, scheduled for launch in the mid-2020s will performexoplanet science via both microlensing surveys and directimaging. Currently, an internal coronagraph is planned toperform starlight suppression for exoplanet imaging. Al-ternatively, an external starshade could be used to achievethe required high contrasts. A starshade could potentiallyhave higher throughput than an equivalent coronagraph,would also have the benefit of a wavelength-independentinner working angle (IWA), and would remove the re-quirement of deformable mirrors in telescope optics. Thestarshade approach would require a separately-launchedocculter spacecraft to be positioned at exact distancesfrom the telescope along the line of sight to a target starsystem. We perform a detailed study to quantify therequirements and feasibility of deploying this additionalspacecraft as a means of exoplanet imaging. The fueluse of the occulter can be categorized into two phases:station-keeping during imaging and repositioning betweendi�erent targets. Target star systems and look vectors aregenerated using Exoplanet Open-Source Imaging Simula-tor (EXOSIMS); a boundary value problem is then solvedbetween successive targets. Given a list of possible targets,the minimum transfers are computed. These methods areused to estimate the number of possible observations thatcould be performed with the expected available fuel masslimit imposed by the likely occulter launch vehicle.

40

Sponsors

41

Cornell GPSAFC

43

book of abstracts...book of abstracts sibley graduate research symposium show us why your research...

Documents