abstract: The rheological properties at liquid-liquid interfaces are important in many industrial processes such as manufacturing foods, pharmaceuticals, cosmetics, and petroleum products. This dissertation focuses on the study of linear viscoelastic properties at liquid-liquid interfaces by tracking the thermal motion of particles confined at the interfaces. The technique of interfacial microrheology is first developed using one- and two-particle tracking, respectively. In one-particle interfacial microrheology, the rheological response at the interface is measured from the motion of individual particles. One-particle interfacial microrheology at polydimethylsiloxane (PDMS) oil-water interfaces depends strongly on the surface chemistry of different tracer particles. In contrast, by tracking the correlated motion of particle pairs, two-particle interfacial microrheology significantly minimizes the effects from tracer particle surface chemistry and particle size. Two-particle interfacial microrheology is further applied to study the linear viscoelastic properties of immiscible polymer-polymer interfaces. The interfacial loss and storage moduli at PDMS-polyethylene glycol (PEG) interfaces are measured over a wide frequency range. The zero-shear interfacial viscosity, estimated from the Cross model, falls between the bulk viscosities of two individual polymers. Surprisingly, the interfacial relaxation time is observed to be an order of magnitude larger than that of the PDMS bulk polymers. To explore the fundamental basis of interfacial nanorheology, molecular dynamics (MD) simulations are employed to investigate the nanoparticle dynamics. The diffusion of single nanoparticles in pure water and low-viscosity PDMS oils is reasonably consistent with the prediction by the Stokes-Einstein equation. To demonstrate the potential of nanorheology based on the motion of nanoparticles, the shear moduli and viscosities of the bulk phases and interfaces are calculated from single-nanoparticle tracking. Finally, the competitive influences of nanoparticles and surfactants on other interfacial properties, such as interfacial thickness and interfacial tension are also studied by MD simulations. / Dissertation/Thesis / Ph.D. Chemical Engineering 2011

abstract: Dye sensitized solar cells (DSSCs) are the third generation solar cells expected to outperform the first two generations of solar cells with their advantages of comparative higher efficiency and lower manufacturing costs. The manufacturing cost of Dye sensitized solar cells is one fifth of the conventional silicon solar cell. However, DSSCs have problems of low conversion efficiency, stability and reliability. Some effective approaches are required to improve their performance. This paper projects the work related to assessment and verification of the repeatability of the semi-automated fabrication process. Changes were introduced in to the fabrication process to enhance the efficiency and stability. The sealant step in the fabrication process was remodeled to a newer version with an improvement in efficiency from 11% to 11.8%. Sputtering was performed on counter electrode in 30 seconds intervals. Cells were fabricated to assess the performance & time dependent characteristics from EIS experiments. Series resistance increased three times in sputtered Pt electrode as compared to standard platinum electrode. This resulted in the degradation of conductive surface on glass electrode due to heavy bombardment of ions. The second phase of the project work relates to the incorporation of SWCNT on the TiO2 electrode and its effect on the cell efficiency. Different weight loadings (0.1 wt %, 0.2 wt%, 0.4 wt %) of SWCNTs were prepared and mixed with the commercial TiO2 paste and ethanol solvent. The TiO2-SWCNT layer was coated on the electrode using screen-printing technique. Both open circuit voltage and photocurrent were found to have measurable dependence on the TiO2 layer loading. Photo voltage ranged from ~0.73 V to ~0.43 V and photocurrent from ~8 to ~33 mA depending on weight percent loading. This behavior is due to aggregation of particles and most TiO2 aggregate particles are not connected to SWCNT. Transparency loss was observed leading to saturation in the photo current and limiting the light absorption within the TiO2 film. / Dissertation/Thesis / M.S.Tech Technology 2011

Dendrite formation on the electrode surface of high energy density batteries, such as lithium (Li) batteries, causes safety problems and limits their applications. Suppressing dendrite growth could significantly improve Li battery performance. In this thesis, computational models are developed to investigate the physics of dendrite formation in Li batteries after nucleation, which strongly depends on the local mass transport. Dendrite growth in various scenarios is studied, such as in an anisotropic electrolyte, a convective electrolyte and structured electrolytes, to understand the effects of mass transport on growth and to investigate mitigation strategies. Various electrolytes lead to different effects on the local mass transport and eventually affect the dendrite morphology in each scenario. Two numerical methods are used in this thesis. The finite difference method (FDM) is adopted to quickly solve the 1D transient mass transport governing equation and the electrostatic Poisson equation. For the more complex 2D reactive mass transport model, the smoothed particle hydrodynamics (SPH) method is employed. The intrinsic advantages of SPH, such as its mesh-free Lagrangian nature, easy implementation of complex physics at the dendrite surface and the well-developed flow modeling capabilities, make it particularly well suited for modeling dendrite growth in the various scenarios studied in this thesis. Based on the results of these computational studies suggestions for improved battery performance are discussed including material properties, such as diffusivity and viscosity of the electrolyte, and cell design improvements such as porosity and tortuosity of a structured electrolyte. These computational studies can help to reduce dendrite growth by suggesting novel battery designs, and play an important part in the development of more stable and reliable high energy density Li batteries. / 2017-02-17T00:00:00Z

This thesis is motivated by time and safety critical applications involving the use of autonomous vehicles to accomplish complex tasks in dynamic and partially known environments. We use temporal logic to formally express such complex tasks. Temporal logic specifications generalize the classical notions of stability and reachability widely studied within the control and hybrid systems communities. Given a model describing the motion of a robotic system in an environment and a formal task specification, the aim is to automatically synthesize a control policy that guarantees the satisfaction of the specification. This thesis presents novel control synthesis algorithms to tackle the problem of motion planning from temporal logic specifications in uncertain environments. For each one of the planning and control synthesis problems addressed in this dissertation, the proposed algorithms are implemented, evaluated, and validated thought experiments and/or simulations. The first part of this thesis focuses on a mobile robot whose success is measured by the completion of temporal logic tasks within a given period of time. In addition to such time constraints, the planning algorithm must also deal with the uncertainty that arises from the changes in the robot's workspace during task execution. In particular, we consider a robot deployed in a partitioned environment subjected to structural changes such as doors that can open and close. The motion of the robot is modeled as a continuous time Markov decision process and the robot's mission is expressed as a Continuous Stochastic Logic (CSL) formula. A complete framework to find a control strategy that satisfies a specification given as a CSL formula is introduced. The second part of this thesis addresses the synthesis of controllers that guarantee the satisfaction of a task specification expressed as a syntactically co-safe Linear Temporal Logic (scLTL) formula. In this case, uncertainty is characterized by the partial knowledge of the robot's environment. Two scenarios are considered. First, a distributed team of robots required to satisfy the specification over a set of service requests occurring at the vertices of a known graph representing the environment is examined. Second, a single agent motion planning problem from the specification over a set of properties known to be satised at the vertices of the known graph environment is studied. In both cases, we exploit the existence of o-the-shelf model checking and runtime verification tools, the efficiency of graph search algorithms, and the efficacy of exploration techniques to solve the motion planning problem constrained by the absence of complete information about the environment. The final part of this thesis extends uncertainty beyond the absence of a complete knowledge of the environment described above by considering a robot equipped with a noisy sensing system. In particular, the robot is tasked with satisfying a scLTL specification over a set of regions of interest known to be present in the environment. In such a case, although the robot is able to measure the properties characterizing such regions of interest, precisely determining the identity of these regions is not feasible. A mixed observability Markov decision process is used to represent the robot's actuation and sensing models. The control synthesis problem from scLTL formulas is then formulated as a maximum probability reachability problem on this model. The integration of dynamic programming, formal methods, and frontier-based exploration tools allow us to derive an algorithm to solve such a reachability problem.

A diagnosis of metastatic cancer reduces a patient's 5-year survival rate by nearly 80% compared to a primary tumor diagnosed at an early stage. While gene expression arrays have revealed unique gene signatures for metastatic cancer cells, we are lacking an understanding of the tangible physical changes that distinguish metastatic tumor cells from each other and from their related primary tumors. At the fundamental level, this translates into first characterizing the phenotype of metastatic cancer cells in vitro both in 2D – looking at morphology and migration – and in 3D – focusing on matrix invasion. While 2D in vitro studies have provided insight into the effects of specific environmental conditions on specific cancer cell lines, the unique details included in each experimental design make it challenging to compare cell phenotype across different in vitro platforms as well as between laboratories and disciplines thatshare the goal of understanding cancer. While 3D phenotype studies have employed more standardized and ubiquitous assays, most available tools lack the imaging capability and geometry to effectively characterize all factors driving 3D matrix invasion. In this work, we present protocols and platforms aimed at addressing the problems identified in the tools currently available for studying metastatic cancer in vitro. First, we present a 2D study of morphology and migration using widely accepted protocols. The study is applied to characterizing phenotypes of three breast cancer cell lines with different metastatic organ tropisms. The results show that general populations of cells from each of the 3 lines are unique in shape and motility despite being derived from the same tumor line and that the observed phenotype differences may be related to differences in focal adhesion assembly. More broadly, these studies suggest that standardizing phenotype studies using commonly available techniques may provide a platform by which to compare phenotypic studies across cancer cell types and between research groups to investigate tropism-specific cancer phenotypes. We conclude our investigation of phenotype with a study of 3D matrix invasion using a novel microfluidic platform. The results show that invasion of metastatic breast cancer cells into a 3D type I collagen gel is significantly enhanced in the presence of live endothelial cells. In applying the model to study cell-cell and cell-matrix interactions driving invasion, our platform revealed that, while the fibronectin-rich matrix deposited by endothelial cells was not sufficient to drive invasion alone, metastatic breast cancer cells were able to exploit a structural or secreted component of energetically inactivated endothelial cell to gain entry into the underlying matrix. These findings have important implications for designing drugs targeted at preventing cancer metastasis. The findings in this dissertation reveal significant phenotypic differences in metastatic breast cancer cells with different preferences in metastatic target organ. In addition, the microfluidic platform reveals novel cell-cell interactions driving a key step in the seeding and colonization of a metastatic tumor. Collectively, these results reveal important characteristics of metastatic cancer cells and their interactions with other cell types during metastasis. These studies also provide platforms on which to target or prevent malignant phenotypes and cellular interactions in the future.

Many computer vision problems boil down to the learning of a good visual similarity function that calculates a score of how likely two instances share the same semantic concept. In this thesis, we focus on two problems related to similarity learning: Person Re-Identification, and Semantic Video Retrieval. Person Re-Identification aims to maintain the identity of an individual in diverse locations through different non-overlapping camera views. Starting with two cameras, we propose a novel visual word co-occurrence based appearance model to measure the similarities between pedestrian images. This model naturally accounts for spatial similarities and variations caused by pose, illumination and configuration changes across camera views. As a generalization to multiple camera views, we introduce the Group Membership Prediction (GMP) problem. The GMP problem involves predicting whether a collection of instances shares the same semantic property. In this context, we propose a novel probability model and introduce latent view-specific and view-shared random variables to jointly account for the view-specific appearance and cross-view similarities among data instances. Our method is tested on various benchmarks demonstrating superior accuracy over state-of-art. Semantic Video Retrieval seeks to match complex activities in a surveillance video to user described queries. In surveillance scenarios with noise and clutter usually present, visual uncertainties introduced by error-prone low-level detectors, classifiers and trackers compose a significant part of the semantic gap between user defined queries and the archive video. To bridge the gap, we propose a novel probabilistic activity localization formulation that incorporates learning of object attributes, between-object relationships, and object re-identification without activity-level training data. Our experiments demonstrate that the introduction of similarity learning components effectively compensate for noise and error in previous stages, and result in preferable performance on both aerial and ground surveillance videos. Considering the computational complexity of our similarity learning models, we attempt to develop a way of training complicated models efficiently while remaining good performance. As a proof-of-concept, we propose training deep neural networks for supervised learning of hash codes. With slight changes in the optimization formulation, we could explore the possibilities of incorporating the training framework for Person Re-Identification and related problems. / 2019-07-09T00:00:00Z

Evaporation from nanopores, owing to its high mass/heat fluxes and high heat transfer coefficients, have found widespread applications in various industrial process, including electronics cooling, solar steam generation, membrane distillation and power generation. To further improve the performance of these nanopore-evaporation-associated processes, it is necessary to experimentally quantify the ultimate transport limit of evaporation from nanopores and understand its dependence on nanoscale confinement and operating conditions. This ultimate transport limit has now been widely accepted to be dictated by evaporation kinetics at the liquid-vapor interface, which is very difficult to quantify experimentally due to the ultra-small evaporation rates from single nanopores. To overcome this challenge, a new measurement approach based on a hybrid nanochannel-nanopore device design has been developed recently. This measurement approach can accurately measure evaporation rates/fluxes from single nanopore and has been used to investigate the effect of nanopore diameter on kinetic-limited evaporation flux. Although this study provides us new fundamental understanding about how nanoscale confinements change evaporation from nanopore, the effects of contaminant and pore porosity, which to some extent determines the practical performance of evaporation from nanopores, have remained elusive. Such lacking understanding has prevented us from developing optimized evaporative nanoporous structures for practical applications. This works aims to investigate the effects of porosity and contaminant on kinetic-limited evaporation flux by experimentally measuring kinetic-limited evaporation rates from nanopore arrays. A modified hybrid nanochannel-nanopore device design is used to achieve this goal. In this modified device design, a nanopore array is directly connected to a 2-D nanochannel and the total evaporation rate from the nanopore array is measured by tracking meniscus receding in the nanochannel during a drying/evaporation process. Using this modified device design, we measured the kinetic-limited evaporation rates from 3x3 nanopore arrays with different interval distances ranging from 200 nm to 1 μm. To facilitate comparison between different devices, the total evaporation rates were converted to evaporation fluxes based on the nanopore projected area. Our results showed that that porosity or nanopore interval distance has negligible effect on the kinetic-limited evaporation flux. We also performed evaporation experiment using water with impurity and studied the effect of contaminant on kinetic-limit evaporation flux. It was observed that the contaminants in water can significantly reduce the kinetic-limited evaporation flux in nanopores and the contaminant effect becomes much more obvious in smaller nanopore due to contaminant-accumulation-induced pore blockage.

This study focuses on scale modelling of a heaving Point Absorber Wave Energy Converter (PAWEC), model verification via wave tank tests and power maximisation control development. Starting from the boundary element method simulation of the wave-PAWEC interaction, linear and non-linear modelling approaches of Wave-To-Excitation-Force (W2EF), Force-To-Motion (F2M), Wave-To-Motion (W2M) are studied. To verify the proposed models, a 1/50 scale PAWEC has been designed, simulated, constructed and tested in a wave tank under a variety of regular and irregular wave conditions. To study the coupling between the PAWEC hydrodynamics and the Power Take-Off (PTO) mechanism, a Finite Element Method (FEM) is applied to simulate and optimise a Tubular Permanent Magnet Linear Generator (TPMLG) as the PTO system and control actuator. Thus linear and non-linear Wave-To-Wire (W2W) models are proposed via combining the W2M and PTO models for the study and development of power maximisation control. The main contributions of this study are summarised as follows: Linear and non-linear F2M models are derived with the radiation force approximated by a finite order state-space model. The non-linear friction is modelled as the Tustin model, a summation of the Stribeck, Coloumb and damping friction forces, whilst the non-linear viscous force is simulated as the drag term in the Morison equation. Thus a non-linear F2M model is derived considering the non-linear friction and viscous forces as a correction or calibration to the linear F2M model. A wide variety of free-decay tests are conducted in the wave tank and the experimental data fit the non-linear F2M modelling results to a high degree. Further, the mechanism how these non-linear factors influence the PAWEC dynamics and energy dissipations is discussed with numerical and experimental results. Three approaches are proposed in this thesis to approximate the wave excitation force: (i) identifying the excitation force from wave elevation, referred to as the W2EF method, (ii) estimating the excitation force from the measurements of pressure, acceleration and displacement, referred to as the Pressure-Acceleration-Displacement-To-Excitation-Force (PAD2EF) approach and (iii) observing the excitation force via an unknown input observer, referred to as the Unknown-Input-Observation-of-Excitation-Force (UIOEF) technique. The W2EF model is integrated with the linear/non-linear F2M models to deduce linear/non-linear W2M models. A series of excitation tests are conducted under regular and irregular wave conditions to verify the W2EF model in both the time- and frequency-domains. The numerical results of the proposed W2EF model show a high accordance to the excitation test data and hence the W2EF method is valid for the 1/50 scale PAWEC. Meanwhile, a wide range of forced-motion tests are conducted to compare the excitation force approximation results between the W2EF, PAD2EF and UIOEF approaches and to verify the linear and non-linear W2M models. Comparison of the PAWEC displacement responses between the linear/non-linear W2M models and forced-motion tests indicates that the non-linear modelling approach considering the friction and viscous forces can give more accurate PAWEC dynamic representation than the linear modelling approach. Based on the 1/50 scale PAWEC dimension and wave-maker conditions, a three-phase TPMLG is designed, simulated and optimised via FEM simulation with special focus on cogging force reduction. The cogging force reduction is achieved by optimise the TPMLG geometric design of the permanent magnets, slots, pole-shoe and back iron. The TPMLG is acting as the PTO mechanism and control actuator. The TPMLG is connected with the buoy rigidly and hence the coupling is achieved by the PTO force. Linear and non-linear W2W models are derived for the study of power maximisation control. To investigate the control performance on the linear and non-linear W2W models, reactive control and phase control by latching are developed numerically with electrical implementation on the TPMLG. Further, a W2W tracking control structure is proposed to achieve power maximisation and displacement constriction under both regular and irregular wave conditions.

abstract: Air pollution is one of the biggest challenges people face today. It is closely related to people's health condition. The agencies set up standards to regulate the air pollution. However, many of the pollutants under the regulation level may still result in adverse health effect. On the other hand, it is not clear the exact mechanism of air pollutants and its health effect. So it is difficult for the health centers to advise people how to prevent the air pollutant related diseases. It is of vital importance for both the agencies and the health centers to have a better understanding of the air pollution. Based on these needs, it is crucial to establish mobile health sensors for personal exposure assessment. Here, two sensing principles are illustrated: the tuning fork platform and the colorimetric platform. Mobile devices based on these principles have been built. The detections of ozone, NOX, carbon monoxide and formaldehyde have been shown. An integrated device of nitrogen dioxide and carbon monoxide is introduced. Fan is used for sample delivery instead pump and valves to reduce the size, cost and power consumption. Finally, the future work is discussed. / Dissertation/Thesis / Ph.D. Electrical Engineering 2012

abstract: The thesis examines how high density polyethylene (HDPE) pipe installed by horizontal directional drilling (HDD) and traditional open trench (OT) construction techniques behave differently in saturated soil conditions typical of river crossings. Design fundamentals for depth of cover are analogous between HDD and OT; however, how the product pipe is situated in the soil medium is vastly different. This distinction in pipe bedding can produce significant differences in the post installation phase. The research was inspired by several incidents involving plastic pipe installed beneath rivers by HDD where the pipeline penetrated the overburden soil and floated to the surface after installation. It was hypothesized that pipes installed by HDD have a larger effective volume due to the presence of low permeability bentonite based drilling fluids in the annular space on completion of the installation. This increased effective volume of the pipe increases the buoyant force of the pipe compared to the same product diameter installed by OT methods, especially in situations where the pipe is installed below the ground water table. To simulate these conditions, a real-scale experiment was constructed to model the behavior of buried pipelines submerged in saturated silty soils. A full factorial design was developed to analyze scenarios with pipe diameters of 50, 75, and 100 mm installed at varying depths in a silty soil simulating an alluvial deposition. Contrary to the experimental hypothesis, pipes installed by OT required a greater depth of cover to prevent pipe floatation than similarly sized pipe installed by HDD. The results suggested that pipes installed by HDD are better suited to survive changing depths of cover. In addition, finite element method (FEM) modeling was conducted to understand soil stress patterns in the soil overburden post-installation. Maximum soil stresses occurring in the soil overburden between post-OT and HDD installation scenarios were compared to understand the pattern of total soil stress incurred by the two construction methods. The results of the analysis showed that OT installation methods triggered a greater total soil stress than HDD installation methods. The annular space in HDD resulted in less soil stress occurring in the soil overburden. Furthermore, the diameter of the HDD annular space influenced the soil stress that occurred in the soil overburden, while the density of drilling fluids did not vastly affect soil stress variations. Thus, the diameter of the annular space could impact soil stress patterns in HDD installations post-construction. With these findings engineers and designers may plan, design, and construct more efficient river-crossing projects. / Dissertation/Thesis / Ph.D. Construction 2012