Esta pesquisa tem como principal objetivo apresentar uma análise do comportamento dos elementos passivos mais utilizados para compensação reativa e na composição dos filtros passivos sintonizados: os capacitores de potência. Tal análise foi focada na comparação de resultados obtidos via simulação computacional com os resultados oriundos de ensaios laboratoriais em ambiente controlado, servindo de base para estabelecer os limites de validade do modelo computacional hoje implementado em softwares comerciais de simulação de fluxo de carga e fluxo de harmônicos de uso bastante difundido no ambiente industrial. Os resultados alcançados demonstram que o modelo clássico de capacitores apresenta uma excelente precisão quando operando em condições ideais. Porém, quando operando na presença de situações adversas às ideais, o modelo demonstrou possuir limitações que serão devidamente reportadas neste documento. / This research has as aim to present a behaviors analysis of the electrical power capacitors which are used for reactive compensation and in the tuned harmonic passive filters. Such analysis was focused on the results comparison obtained via computer simulation with the results obtained in laboratory tests done in a controlled environment, providing the basis to establish the validity limits of the computational model implemented in today\'s commercial load flow and harmonic flow software simulation widespread use in industrial environments. Results show that the classical model capacitor has excellent accuracy when operating under ideal conditions. However, when operating in presence of adverse situations the classical model has demonstrated limitations that will be duly and reported in this work.

Existe atualmente na sociedade um grande número de sistemas reais de alta complexidade. Esta complexidade pode ser definida tanto do ponto de vista da dificuldade em identificar todas as partes que compõem estes sistemas, como também, pela compreensão e definição real da relação entre estas partes, permitindo, desta forma uma representação adequada do comportamento global do sistema. O comportamento global destes sistemas não se caracteriza pela soma do comportamento de suas partes componentes. Normalmente, a modelagem destes sistemas não reflete, de forma realística, o seu comportamento, devido ao excesso de simplificações realizadas. Por outro lado, alguns modelos são impraticáveis de serem aplicados, devido ao excessivo esforço computacional e a restrições de tempo. O presente trabalho de pesquisa apresenta uma proposta de um modelo de otimização para um problema real de alta complexidade e com fortes requisitos de segurança (safety) encontrado na Infra-estrutura Aeronáutica Brasileira e Mundial. Este problema está relacionado ao desbalanceamento entre a capacidade e demanda em infra-estrutura aeronáutica em sistemas de transporte aéreo. Para tanto, o trabalho propõe um Modelo de Otimização de Demanda (MOD) em infra-estrutura aeronáutica, através da técnica de Inteligência Artificial denominada de Algoritmos Genéticos. A pesquisa analisa a eficiência do modelo proposto em termos da resolução do problema, bem como quanto à qualidade das respostas apresentadas. De forma complementar é avaliada a importância de cada um dos parâmetros do modelo de otimização através da sua flexibilização. / Nowadays, in the society, there are a great number of real systems with high complexity. This complexity can be justified in function of the difficulty in identifying all parts that compose these systems, but also the complex relationship between them. The global behavior of these systems is not characterized for the addition of the behavior of its contracting parties. Normally, the modeling of these systems does not reflect its realistic behavior, due the excess of simplifications carried out. On the other hand, some models are impracticable to be solved, because the extreme computational effort necessary. The present research develops a proposal of an optimization model for a real problem of high complexity and with hard safety requirements found in the Brazilian and world-wide aeronautical infrastructure. This problem deals with the unbalancing between the capacity and demand in infrastructure aeronautics in air transportation systems. The work considers a Demand Optimization Model (DOM) for aeronautical infrastructure through the technique of artificial intelligence denominated Genetic Algorithms. The research analyzes the efficiency of the considered model in terms of problem resolution, as well as, the quality of the presented answers. Of complementary form, some parameters of the model were adjusted and their importance were avaluated.

O interesse no uso de técnicas de aquisição de dados 3D em vídeos interativos e sistemas de videoconferência imersiva aumentou significativamente nos últimos anos. A disponibilidade de sensores de profundidade baseados em luz estruturada no varejo apresenta uma série de novas oportunidades em sistemas de teleconferência, entretenimento e e-learning. No entanto, problemas de oclusão frequentemente ocorrem quando é utilizado um único sensor ou técnicas de reconstrução 2,5D. Outras técnicas de interação e aplicações de teleconferências podem exigir uma reconstrução mais completa do usuário. Uma possível solução para isso implica na utilização de múltiplos dispositivos simultaneamente, o que traz novos desafios relacionados ao processo de calibração de câmaras e a aferição do aumento da superfície capturada quando é adicionado um novo sensor. Neste trabalho é apresentada a proposta VMD (Video-avatar from Multiple Depth maps), uma arquitetura de sistema que trata esses dois problemas. Com a finalidade de validar a viabilidade dessa proposta, três experimentos foram feitos resultando em uma análise qualitativa e quantitativa a fim de encontrar uma configuração adequada de múltiplos sensores de profundidade. / The interest in the use of 3D data acquisition techniques in video based interaction and immersive videoconferencing systems has noticeably increased in recent years. The availability of structured light 3D sensors as off-the-shelf hardware presents a series of new opportunities in teleconferences, entertainment and e-learning. However, occlusion issues often occur when it is used a single sensor or 2,5D reconstruction techniques. Other interaction techniques and teleconferencing applications may require a more complete reconstruction of the user. One possible solution to this involves the use of multiple devices simultaneously, which may bring new challenges related to the calibration process of cameras and measurement of the increase in covered area when it is added a new sensor. In this work we present VMD (Video-avatar from Multiple Depth maps), a system architecture that addresses this two issues. In order to validate the feasibility of this proposal, three experiments were performed resulting in a qualitative and quantitative analysis with the purpose of finding a suitable setup with multiple depth sensors.

Muitos componentes de máquinas, equipamentos e meios de transporte podem ser representados na forma de corpos rombudos cujos deslocamentos em meio fluido podem gerar ruído. Nesse contexto, a tese tem como proposta o estudo da previsão numérica do ruído gerado oriundo do escoamento em torno de corpos rombudos. Como exemplo representativo de um corpo rombudo, o som gerado do escoamento ao redor de um cilindro é estudado. O escoamento em torno de um cilindro é considerado com esteira tridimensional e turbulenta na condição do regime subcrítico, que é caracterizado pela separação da camada limite no regime laminar. O escoamento em torno de um cilindro é obtido através da simulação numérica não-estacionária considerando domínio computacional tridimensional. Para lidar com a turbulência e a tridimensionalidade, o escoamento é resolvido utilizando a metodologia de simulação de grandes escalas (LES). O domínio computacional é discretizado pelo método de volumes finitos. O ruído é calculado separadamente utilizando a analogia de Ffwocs Williams & Hawkings (FW-H), cuja equação de onda tem como termo fonte a solução do escoamento fornecida pela simulação numérica. Na analogia de FW-H, a flutuação de pressão acústica é obtida no campo afastado assumindo um meio quiescente entre a região das fontes sonoras (campo próximo) e o local considerado para o cálculo do ruído. Devido ao alto custo computacional da simulação tridimensional (3D), as simulações numéricas foram realizadas com comprimento de envergadura do cilindro limitado, possibilitando considerar parte dos efeitos tridimensionais da esteira. No cálculo final do ruído, os métodos de correção acústica de Kato et al. (1993) e Seo & Moon (2007) são empregados de forma a equiparar a geração sonora obtida com cilindro de menor comprimento de envergadura ao respectivo ruído obtido experimentalmente com cilindro de maior comprimento. Esta tese contribuiu com uma investigação numérica da metodologia de cálculo de ruído utilizando a analogia de Ffwocs Williams & Hawkings para um escoamento tridimensional em torno de um cilindro considerando número de Reynolds elevado de Re = 90.000 e Re = 22.000. Os resultados mostram que a metodologia é capaz de prever o som no campo afastado nos casos simulados, visto que os espectros sonoros obtidos concordam com os respectivos dados experimentais. Além disso, os métodos de correção acústica de Kato et al. (1993) e Seo & Moon (2007) mostraram ser adequados para o cálculo do ruído adicional, que considera um cilindro de comprimento maior, porém são dependentes da correta estimativa do comprimento de coerência do escoamento ao longo da envergadura do cilindro. Por fim, o espectro sonoro obtido no campo afastado é função do resultado fornecido pela simulação numérica do escoamento, principalmente quanto à coerência entre as flutuações de pressão ao longo do comprimento de envergadura do cilindro e à amplitude de flutuação de força na direção da sustentação exercida na parede do cilindro. / Many components of machines, equipments and means of transport can be represented as a bluff body whose motion through a fluid can generate noise. In this context, this thesis is focused on the study of numerical prediction of noise generated by the flow around bluff bodies. As an example of bluff body, the sound generated from flow around a circular cylinder is studied. The flow over a circular cylinder is investigated by considering the wake as tridimensional and turbulent in the subcritical regime, which is characterized by a laminar boundary layer separation. The flow over a circular cylinder is obtained by time-dependent numerical simulation considering three-dimensional computational domain. In order to cope with turbulence and three-dimensionality, the flow is solved using the Large Eddy Simulation (LES) methodology. The computational domain is discretized by the finite volume method. The noise is calculated separately using the Ffwocs Williams & Hawkings (FW-H) analogy, whose wave equation has as a source term the flow solution provided by the numerical simulation. With regard to the FW-H analogy, the acoustic pressure fluctuation is obtained in the far-field by assuming a quiescent medium between the sound sources region (near-field) and the location considered for acoustic computation. Due to the high computational cost of three-dimensional (3D) simulation, the numerical simulations were conducted with a cylinder span length limited in size, which allows taking into account part of the wake three-dimensionality. Regarding the final acoustic computation, the acoustic correction methods of Kato et al. (1993) and Seo & Moon (2007) are used in order to match the sound obtained by the short cylinder span to the correspondent sound obtained experimentally for a long cylinder span. This thesis contributed to investigate numerically the computational method of applying the Ffwocs Williams & Hawkings analogy for solving the noise generated from a threedimensional flow over a circular cylinder with high Reynolds number, particularly, at Re = 90,000 and Re = 22,000. The results show that this computational method is able to predict the far-field sound for the simulated cases, since the noise spectra obtained are found to be in agreement with the corresponding experimental data. In addition, the acoustic correction method of Kato et al. (1993) and Seo & Moon (2007) provided good predictions with regard to the adding noise computation, however, its results are dependent on accurate estimation of the spanwise coherence length of the flow. Finally, the sound spectrum obtained in the far-field is tied to the flow behavior provided by the numerical simulation, especially regarding the coherence between the pressure fluctuations over the spanwise length of the cylinder and the lift force fluctuation amplitude exerted on the cylinder wall.

A adubação tem o intuito de fornecer nutrientes (solutos) às plantas em quantidades requeridas no momento adequado, visando obter produções agrícolas economicamente viáveis. Todavia, em alguns casos, o manejo da adubação não é feito de maneira correta, principalmente por se utilizar quantidades excessivas desses nutrientes. Juntando-se a esse fato, a água da irrigação e/ou da chuva, pode vir a ser um facilitador do escoamento superficial e da lixiviação através do perfil de solo, tornando-se, dessa forma, indisponível às plantas e, podem provocar a poluição de águas superficiais através da erosão hídrica e eutrofização, além da contaminação de águas subterrâneas e a acidificação de camadas do solo. Portanto, para evitar esses impactos ambientais, é necessário conhecer e entender a dinâmica de solutos no perfil de solo mediante a obtenção de informações que venham a caracterizar esses processos. Dentro dessas informações, tem-se a obtenção dos parâmetros de transporte de solutos. Tais parâmetros podem sofrer várias interferências do meio poroso, dentre elas a composição físico-química e mineralógica dos solos, a qual poderá determinar a capacidade de retenção ou dispersão desses íons. Dessa forma, o objetivo dessa pesquisa foi avaliar o efeito da composição físico-química e mineralógica de alguns tipos de solo sobre a dinâmica dos íons nitrato, fósforo e potássio, mediante a análise de seus parâmetros de transporte, obtidos pelo ajuste numérico das curvas de distribuição de efluentes (Breakthrough Curves, BTC) aplicando o código CFITIM dentro do software STANMOD (STudio of ANalytical MODels). Além disso, utilizando-se colunas com diferentes alturas (10 e 20 cm) e buscou-se recomendar qual altura de coluna possui acurácia sobre os parâmetros mediante simulações numéricas das BTC\'s pelo modelo HYDRUS-1D. A pesquisa foi conduzida junto ao Departamento de Engenharia de Biossistemas da Universidade de São Paulo (USP/ESALQ). Os tratamentos foram distribuídos em um esquema fatorial de 5 x 2, ou seja, cinco tipos de solo e duas alturas de colunas de solo para cada íon, individualmente. O delineamento experimental foi em blocos casualizados, com três blocos, totalizando 30 parcelas por soluto. Para a análise dos resultados, os parâmetros de transportes foram submetidos à análise de variância e a avaliação do desempenho do modelo pelos índices estatísticos: Raiz do Erro Quadrático Médio (RMSE), Erro Absoluto Médio (MAE) e o Coeficiente de Determinação (R2). Diante dos resultados obtidos, o solo PV foi diferente estatisticamente dos demais solos em todos os íons no coeficiente de dispersão. Diante das diferentes nas alturas das colunas, os solos PV, PVAe, NV e PVAd apresentaram diferenças significativas no fator de retardamento e os coeficientes de distribuição e dispersão nos íons nitrato e potássio. Sendo assim, recomendou-se a altura de 20 cm, pois os parâmetros determinados foram precisos em relação a 10 cm, pelo fato do modelo HYDRUS-1D ter alcançado um desempenho adequado nas simulações das BTC\'s nos solos arenosos e argilosos. / Fertilizer is applied to provide nutrients (solutes) for plants in required quantities, and at the right time, to facilitate economically viable agricultural production. However, in some cases excessive quantities of fertilizers are applied. Such overapplication can result in runoff and ions leaching through the soil profile, where they are unavailable to plants, and can cause surface water pollution by water erosion and eutrophication, besides of groundwater contamination and soil acidification. To avoid these environmental impacts, it is necessary to characterize the transport parameters that govern the dynamic of solutes in soil profile. These parameters are influenced by porous media properties, such as mineralogical composition of soils, that regulate the retention or dispersion of ions. The objective of this study was to evaluate the effect of physicochemical and mineralogical composition of soils on transport parameters of nitrate, phosphorus, and potassium, estimated by using the CFITIM code inside of STANMOD (STudio of ANalytical MODels) software to fit BTC\'s. Besides that, using of columns with different heights (10 and 20 cm) and recommended the columns height with accuracy under transport parameters by numerical simulations of the BTC\'s using the HYDRUS-1D model. The research was conducted in the Department of Biosystems Engineering, University of São Paulo (USP/ESALQ). The treatments were a 5 x 2 factorial with five soils and two column heights for each individual solute. The experimental design was random blocks, with three blocks, totaling 30 plots for each solute. For analysis of results, the transport parameters were submitted a variance analysis and the evaluation of model performance by statistical indexes: Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Determination Coefficient (R2). The dispersion coefficient for PV soil were significantly different from those for other soils, for all ions. In front of the different heights of the columns, PV, PVAe, NV and PVAd soils showed significant differences in retardation factor, and distribution and dispersion coefficients, for nitrate and potassium ions. The parameters obtained from 20 cm columns were accurate than those obtained from 10 cm height, based on adequate performance of HYDRUS-1D model in BTC\'s simulations of sandy and clay soils.

The environment of knowledge exchange between computation and biology elicits a contemporary approach towards architecture. Computation, as an overarching mode of thinking, instructs the analysis, understanding and reinterpretation of the un-formal structure of natural organizations (such as systematic construct, information flow, and process through time) for architectural form generation. Consequently, the computing theory originates a mind-shift where processes, relations, and dependencies are a major concern for reconsidering and re-comprehending the environment. Besides, computation presents universal modes of thinking and tools for modeling, within which transdisciplinary studies and knowledge interchange between distinct disciplines are flourished. This thesis will discuss architectural form generation through interpreting computation as &ldquo / transcoding&rdquo / and an interface, while nature will be regarded as a &ldquo / model&rdquo / and a source for learning. A case study will be conducted by analyzing cactus plants and their common generative logic in the framework of computation. Consequently, the produced computational model of cactus plants will be scrutinized for probable outcomes, questioning what such a re-interpretation of natural systems may imply for architecture.

The dynamics of biological neural networks are of great interest to neuroscientists and are frequently studied using conductance-based compartmental neuron models. For speed and ease of use, neuron models are often reduced in morphological complexity. This reduction may affect input processing and prevent the accurate reproduction of neural dynamics. However, such effects are not yet well understood. Therefore, for my first aim I analyzed the processing capabilities of 'branched' or 'unbranched' reduced models by collapsing the dendritic tree of a morphologically realistic 'full' globus pallidus neuron model while maintaining all other model parameters. Branched models maintained the original detailed branching structure of the full model while the unbranched models did not. I found that full model responses to somatic inputs were generally preserved by both types of reduced model but that branched reduced models were better able to maintain responses to dendritic inputs. However, inputs that caused dendritic sodium spikes, for instance, could not be accurately reproduced by any reduced model. Based on my analyses, I provide recommendations on how to construct reduced models and indicate suitable applications for different levels of reduction. In particular, I recommend that unbranched reduced models be used for fast searches of parameter space given somatic input output data. The intrinsic electrical properties of neurons depend on the modifiable behavior of their ion channels. Obtaining a quality match between recorded voltage traces and the output of a conductance based compartmental neuron model depends on accurate estimates of the kinetic parameters of the channels in the biological neuron. Indeed, mismatches in channel kinetics may be detectable as failures to match somatic neural recordings when tuning model conductance densities. In my first aim, I showed that this is a task for which unbranched reduced models are ideally suited. Therefore, for my second aim I optimized unbranched reduced model parameters to match three experimentally characterized globus pallidus neurons by performing two stages of automated searches. In the first stage, I set conductance densities free and found that even the best matches to experimental data exhibited unavoidable problems. I hypothesized that these mismatches were due to limitations in channel model kinetics. To test this hypothesis, I performed a second stage of searches with free channel kinetics and observed decreases in the mismatches from the first stage. Additionally, some kinetic parameters consistently shifted to new values in multiple cells, suggesting the possibility for tailored improvements to channel models. Given my results and the potential for cell specific modulation of channel kinetics, I recommend that experimental kinetic data be considered as a starting point rather than as a gold standard for the development of neuron models.

Ongoing advances in tissue engineering with the goal to address the clinical shortage of donor organs have encouraged the design and development of biomaterials to be used in tissue-engineered scaffolds. Furthermore, biomaterials have been used as delivery vehicles for vaccines that aim to enhance the protective immunity against pathogenic agents. These tissue-engineered constructs or vaccines are usually combination products that combine biomaterial and biological (e.g. cells, proteins, and/or DNA) components. Upon introduction into the body, the host response towards these products will be a combination of both a non-specific inflammatory response towards the biomaterial and an antigen-specific immune response towards the biological component(s). Recently, the biomaterial component was shown to influence the immune response towards a co-delivered antigen. Specifically, poly(lactic-co-glycolic acid) (PLGA), but not agarose, scaffolds or microparticles (MPs) enhanced the humoral response to a model antigen, ovalbumin. This in vivo result echoed with the in vitro study that PLGA, but not agarose, supported a mature phenotype of dendritic cells (DCs), the most potent antigen-presenting cells. Therefore, it is hypothesized that the effect of biomaterials on DC phenotype may influence the adaptive immunity against a co-delivered antigen. Understanding how biomaterials affect DC response will facilitate the selection and design of biomaterials that direct a desired immune response for tissue engineering or vaccine delivery applications. The objectives of this research were to elucidate the correlations between material properties and DC phenotype, develop predictive models for DC response based on material properties, and uncover the molecular basis for DC response to biomaterials. Well-defined biomaterial systems, including clinical titanium (Ti) substrates and two polymer libraries, were chosen to study induced DC phenotype. Due to the time-consuming nature of conventional methods for assessing DC phenotype, a high-throughput (HTP) method was first developed to screen for DC maturation based on surface marker expression (CHAPTER 4). A 96-well filter plate-based HTP methodology was developed and validated for the assessment of DC response to biomaterials. A "maturation factor", defined as CD86/DC-SIGN and measured by immunostaining, was found to be a cell number-independent metric for DC maturation and could be adapted to screen for DC maturation in a microplate format. This methodology was shown to reproducibly yield similar results of DC maturation in response to biomaterial treatment as compared to the conventional flow cytometric method upon DC treatment in 6-well plates. In addition, the supernatants from each treatment could easily be collected for cytotoxicity assessment using glucose-6-phosphate dehydrogenase (G6PD)-based assay and cytokine profiling using multiplex technology. In other words, the 96-well filter plate-based methodology can generate three outcomes from one single cell culture: 1) maturation marker expression, 2) cytotoxicity, and 3) cytokine profile. To examine which material properties were critical in determining DC phenotype, a set of three clinical titanium (Ti) substrates with well-defined surfaces was used to treat DCs (CHAPTER 5). These Ti substrates included pretreatment (PT), sand-blasted and acid-etched (SLA), and modified SLA (modSLA), with different roughness and surface energy. DCs responded differentially to these substrates. Specifically, PT and SLA induced a mature DC (mDC) phenotype, while modSLA-treated DCs remained immature based on surface marker expression, cytokine production profiles and cell morphology. Both PT and SLA induced higher CD86 expression as compared to iDC control, while modSLA maintained CD86 expression at a level similar to iDC. PT- or SLA-treated DCs exhibited dendritic processes associated with a mDC phenotype, while modSLA-treated DCs were rounded, a morphology associated with an iDC phenotype. Furthermore, PT induced increased secretion of MCP-1 by DCs compared to iDCs, indicating that PT promoted a pro-inflammatory environment. SLA induced higher IL-16 production, which is a pleiotropic cytokine, by DCs, most likely as a pro-inflammatory response due to the enhanced maturation of DCs induced by SLA. In contrast, modSLA did not induced enhanced production of any cytokines examined. Principal component analysis (PCA) were used to reduce the multi-dimensional data space and confirmed these experimental results, and it also indicated that the non-stimulating property of modSLA co-varied with certain surface properties, such as high surface hydrophilicity, % oxygen and % titanium of the substrates. In contrast, high surface % carbon and % nitrogen were more associated with a mDC phenotype. Furthermore, PCA also suggested that surface line roughness (Ra) did not contribute to the expression of CD86, an important maturation marker, suggesting that roughness had little impact on DC response (CHAPTER 5). DC response-material property relationships were also derived using more complex materials from a combinatorial library of polymethacrylates (pMAs) (CHAPTER 6). Twelve pMAs were selected and were found to induce differential DC response using the HTP method described in CHAPTER 4. These pMAs resulted in a trend of increasing DC maturation represented by the metric CD86/DC-SIGN, which was consistent with the trends of the production of pro-inflammatory cytokine, TNF-α, and chemokine, IL-8. Interestingly, this set of pMAs induced an opposite trend of IL-16 production, which is most likely released as an anti-inflammatory cytokine in this situation. These polymers were characterized extensively for a number of material properties, including surface chemical composition, glass transition temperature (Tg), air-water contact angle, line roughness (Ra), surface roughness (Sa), and surface area. Similar to the results from the Ti study, PCA determined that surface carbon correlated with enhanced DC maturation, while surface oxygen was associated with an iDC phenotype. In addition, Tg, Ra, and surface area were unimportant in determining DC response. Partial square linear regression (PLSR), a multivariate modeling approach, was implemented using the pMAs as the training set and a separate polymer library, which contained methacrylate- and acrylate-based terpolymers, as the prediction set. This model successfully predicted DC phenotype in terms of surface marker expression with R2prediction = 0.76. Furthermore, prediction of DC phenotype was effective based on only theoretical chemical composition of the bulk polymers with R2prediction = 0.80 (CHAPTER 6). Nonetheless, one should note that a predictive model can be only as good as what it is trained on and cannot be used to predict the DC response induced by a type of materials different from the training set. Also, this model might not contain all the important material properties such as polymer swelling and cannot predict specific types of immune responses. However, these results demonstrated that a generalized immune cell response can be predicted from biomaterial properties, and computational models will expedite future biomaterial design and selection (CHAPTER 6). From the pMA library, pMAs that induced the two extremes of DC phenotype (mature or immature) were identified for elucidating the mechanistic basis of biomaterial-induced DC responses (CHAPTER 7). Two pMAs, polyhydroxyethylmethacrylate (pHEMA) and poly(isobutyl-co-benzyl-co-terahydrofurfuryl)methacrylate (pIBTMA), were selected because they induced the least and the most mature DC phenotype, respectively. These pMAs were used to elucidate the activation profiles of transcription factors in DCs after biomaterial treatment and were compared to the iDC and mDC controls. In addition, a combined treatment of pHEMA and LPS was also included to determine if pHEMA could maintain an iDC phenotype in the presence of LPS. Interestingly, pIBTMA induced DC maturation primarily through the activation of NF-κB, while pHEMA mediated suppression of DC maturation through multiple TFs, including the activation of ISRE, E2F-1, GR-PR, NFAT, and HSF. GR-PR and E2F-1 have been shown to be associated with the suppression of DC maturation; ISRE, E2F-1, and NFAT are linked to apoptosis induction; HSF regulates the production of heat shock proteins (HSPs) that induce DC maturation and inhibit apoptosis. The activation of HSF by pHEMA was most likely a natural defensive mechanism against the other apoptotic signals. Therefore, pHEMA suppressed DC maturation through the induction of apoptosis. Surprisingly, in the presence of pHEMA, the effect of LPS was completely eliminated, suggesting that biomaterials can override the effect of soluble factors. The morphology and surface marker expression of DCs treated with these different biomaterials or controls were consistent with TF activation profiles (CHAPTER 7). Overall, this research illustrates that biomaterial properties, within the chosen biomaterial space, can be correlated to DC phenotype and more importantly, can be used as predictors for relative levels of DC phenotype. Furthermore, the differential responses induced by different biomaterials were mediated through the distinct activation profiles of transcription factors. Together, these findings are expected to facilitate the design and selection of biomaterials that direct desired immune responses.

With its long cycle life and scalable design, the vanadium redox flow battery (VRB) is a promising technology for grid energy storage. However, high materials costs have impeded its commercialization. An essential but costly component of the VRB is the ion-exchange membrane. The ideal VRB membrane provides a highly conductive path for protons, prevents crossover of reactive species, and is tolerant of the acidic and oxidizing chemical environment of the cell. In order to study membrane performance and optimize cell design, mathematical models of the separator membrane have been developed. Where previous VRB membrane models considered minimal details of membrane transport, generally focusing on conductivity or self-discharge at zero current, the model presented here considers coupled interactions between each of the major species by way of rigorous material balances and concentrated solution theory. The model describes uptake and transport of sulfuric acid, water, and vanadium ions in Nafion membranes, focusing on operation at high current density. Governing equations for membrane transport are solved in finite difference form using the Newton-Raphson method. Model capabilities were explored, leading to predictions of Ohmic losses, vanadium crossover, and electro-osmotic drag. Experimental methods were presented for validating the model and for further improving estimates of uptake parameters and transport coefficients. / text

Eye movements are a powerful tool to examine cognitive processes. However, in most paradigms little is known about the dynamics present in sequences of saccades and fixations. In particular, the control of fixation durations has been widely neglected in most tasks. As a notable exception, both spatial and temporal aspects of eye-movement control have been thoroughly investigated during reading. There, the scientific discourse was dominated by three controversies, (i), the role of oculomotor vs. cognitive processing on eye-movement control, (ii) the serial vs. parallel processing of words, and, (iii), the control of fixation durations. The main purpose of this thesis was to investigate eye movements in tasks that require sequences of fixations and saccades. While reading phenomena served as a starting point, we examined eye guidance in non-reading tasks with the aim to identify general principles of eye-movement control. In addition, the investigation of eye movements in non-reading tasks helped refine our knowledge about eye-movement control during reading. Our approach included the investigation of eye movements in non-reading experiments as well as the evaluation and development of computational models. I present three main results : First, oculomotor phenomena during reading can also be observed in non-reading tasks (Chapter 2 & 4). Oculomotor processes determine the fixation position within an object. The fixation position, in turn, modulates both the next saccade target and the current fixation duration. Second, predicitions of eye-movement models based on sequential attention shifts were falsified (Chapter 3). In fact, our results suggest that distributed processing of multiple objects forms the basis of eye-movement control. Third, fixation durations are under asymmetric control (Chapter 4). While increasing processing demands immediately prolong fixation durations, decreasing processing demands reduce fixation durations only with a temporal delay. We propose a computational model ICAT to account for asymmetric control. In this model, an autonomous timer initiates saccades after random time intervals independent of ongoing processing. However, processing demands that are higher than expected inhibit the execution of the next saccade and, thereby, prolong the current fixation. On the other hand, lower processing demands will not affect the duration before the next saccade is executed. Since the autonomous timer adjusts to expected processing demands from fixation to fixation, a decrease in processing demands may lead to a temporally delayed reduction of fixation durations. In an extended version of ICAT, we evaluated its performance while simulating both temporal and spatial aspects of eye-movement control. The eye-movement phenomena investigated in this thesis have now been observed in a number of different tasks, which suggests that they represent general principles of eye guidance. I propose that distributed processing of the visual input forms the basis of eye-movement control, while fixation durations are controlled by the principles outlined in ICAT. In addition, oculomotor control contributes considerably to the variability observed in eye movements. Interpretations for the relation between eye movements and cognition strongly benefit from a precise understanding of this interplay. / Blickbewegungen stellen ein wichtiges Instrument dar, um kognitive Prozesse zu untersuchen. In den meisten Paradigmen ist allerdings wenig über die Entstehung von Sakkaden und Fixationen bekannt. Insbesondere die Kontrolle der Fixationsdauern wurde häufig außer acht gelassen. Eine wesentliche Ausnahme stellt die Leseforschung dar, in der sowohl zeitlichliche als auch räumliche Aspekte der Blickbewegungssteuerung im Detail betrachtet wurden. Dabei war der wissenschaftliche Diskurs durch drei Kontroversen gekennzeichnet, die untersuchten, (i), welchen Einfluss okulomotorische bzw. kognitive Prozesse auf die Blicksteuerung haben, (ii), ob Worte seriell oder parallel verarbeitet werden und, (iii), wie Fixationsdauern kontrolliert werden. Die vorliegende Arbeit zielt im wesentlichen darauf ab, die Dynamik von Fixationssequenzen zu erforschen. Ausgehend von den Erkenntnissen beim Lesen untersuchten wir Blickbewegungen in Nichtlese-Aufgaben, mit dem Ziel allgemeine Prinzipien der Blicksteuerung zu identifizieren. Zusätzlich versuchten wir mit Hilfe dieser Aufgaben, Erkenntnisse über Prozesse beim Lesen zu vertiefen. Unser Vorgehen war sowohl von der Durchführung von Experimenten als auch der Entwicklung und Evaluation computationaler Modelle geprägt. Die Hauptbefunde zeigten: Erstens, okulomotorische Phänomene des Lesens lassen sich in Suchaufgaben ohne Wortmaterial replizieren (Kapitel 2 & 4). Dabei bestimmen okulomotorische Prozesse die Fixationsposition innerhalb eines Objektes. Diese wiederum beeinflusst das nächste Sakkadenziel sowie die Fixationsdauer. Zweitens, wesentliche Vorhersagen von Modellen, in denen Blickbewegungen von seriellen Aufmerksamkeitsverschiebungen abhängen, konnten falsifiziert werden (Kapitel 3). Stattdessen legen unsere Erkenntnisse nahe, dass die Blicksteuerung von der parallelen Verarbeitung mehrerer Objekte abhängt. Drittens, Fixationsdauern werden asymmetrisch kontrolliert (Kapitel 4). Während hohe Verarbeitungsanforderungen Fixationsdauern unmittelbar verlängern können, führen niedrige Verarbeitungsanforderungen nur zeitlich verzögert zu einer Reduktion. Wir schlagen ein computationales Modell ICAT vor, um asymmetrische Kontrolle zu erklären. Grundlage des Modells ist ein autonomer Zeitgeber, der unabhängig von der momentanen Verarbeitung nach zufälligen Zeitintervallen Sakkaden initiiert. Unerwartet hohe Verarbeitungsanforderungen können die Initiierung der nächsten Sakkade hinauszögern, während unerwartet niedrige Verarbeitungsanforderungen den Beginn der nächsten Sakkade nicht verändern. Der Zeitgeber passt sich allerdings von Fixation zu Fixation neuen Verarbeitungsanforderungen an, so dass es zu einer zeitlich verzögerten Reduktion der Fixationsdauern kommen kann. In einer erweiterten Version des Modells überprüfen wir die Kompatibilität ICATs mit einer realistischen räumlichen Blicksteuerung. Die Ähnlichkeit von Blickbewegungsphänomenen über Aufgaben hinweg legt nahe, dass sie auf allgemeinen Prinzipien basieren. Grundlage der Blicksteuerung ist die verteilte Verarbeitung des visuellen Inputs, während die Kontrolle der Fixationsdauer auf den Prinzipien von ICAT beruht. Darüber hinaus tragen okulomotorische Phänomene wesentlich zur Variabilität der Blicksteuerung bei. Ein Verständnis dieses Zusammenspiels hilft entscheidend den Zusammenhang von Blickbewegungen und Kognitionen besser zu verstehen.