Improvements in performance, reliability and durability as well as reductions in production costs, remain critical prerequisites for the commercialization of proton exchange membrane fuel cells. In this thesis, a computational framework for fuel cell analysis and optimization is presented as an innovative alternative to the time consuming trial-and-error process currently used for fuel cell design. The framework is based on a two-dimensional through-the-channel isothermal, isobaric and single phase membrane electrode assembly (MEA) model. The model input parameters are the manufacturing parameters used to build the MEA: platinum loading, platinum to carbon ratio, electrolyte content and gas diffusion layer porosity. The governing equations of the fuel cell model are solved using Netwon's algorithm and an adaptive finite element method in order to achieve quadratic convergence and a mesh independent solution respectively. The analysis module is used to solve two optimization problems: i) maximize performance; and, ii) maximize performance while minimizing the production cost of the MEA. To solve these problems a gradient-based optimization algorithm is used in conjunction with analytical sensitivities. The presented computational framework is the first attempt in the literature to combine highly efficient analysis and optimization methods to perform optimization in order to tackle large-scale problems. The framework presented is capable of solving a complete MEA optimization problem with state-of-the-art electrode models in approximately 30 minutes. The optimization results show that it is possible to achieve Pt-specific power density for the optimized MEAs of 0.422 $g_{Pt}/kW$. This value is extremely close to the target of 0.4 $g_{Pt}/kW$ for large-scale implementation and demonstrate the potential of using numerical optimization for fuel cell design.

Improvements in performance, reliability and durability as well as reductions in production costs, remain critical prerequisites for the commercialization of proton exchange membrane fuel cells. In this thesis, a computational framework for fuel cell analysis and optimization is presented as an innovative alternative to the time consuming trial-and-error process currently used for fuel cell design. The framework is based on a two-dimensional through-the-channel isothermal, isobaric and single phase membrane electrode assembly (MEA) model. The model input parameters are the manufacturing parameters used to build the MEA: platinum loading, platinum to carbon ratio, electrolyte content and gas diffusion layer porosity. The governing equations of the fuel cell model are solved using Netwon's algorithm and an adaptive finite element method in order to achieve quadratic convergence and a mesh independent solution respectively. The analysis module is used to solve two optimization problems: i) maximize performance; and, ii) maximize performance while minimizing the production cost of the MEA. To solve these problems a gradient-based optimization algorithm is used in conjunction with analytical sensitivities. The presented computational framework is the first attempt in the literature to combine highly efficient analysis and optimization methods to perform optimization in order to tackle large-scale problems. The framework presented is capable of solving a complete MEA optimization problem with state-of-the-art electrode models in approximately 30 minutes. The optimization results show that it is possible to achieve Pt-specific power density for the optimized MEAs of 0.422 $g_{Pt}/kW$. This value is extremely close to the target of 0.4 $g_{Pt}/kW$ for large-scale implementation and demonstrate the potential of using numerical optimization for fuel cell design.

The continuous flow separation of magnetic particles from a mixture of particles could improve the performance of magnetic bead based assays but the formation of agglomerates limit the separation efficiency. Bead agglomerates are formed as a result of magnetic binding forces while the hydrodynamic fluid environment strongly influences their movement. The ability to predict the interaction between nearby beads will help to determine a threshold separation distance which will be recommended for use when obtaining measurement within a magnetic bead assay for a specified time interval. The introductory part of this thesis explored the development of a two dimensional numerical model in Matlab which predicts the trajectory pattern as well as magnetic induced velocities between a pair of super-paramagnetic beads suspended in water within a uniform field. The movement of a bead pair interacting due to both magnetic and hydrodynamic forces within a magnetophoretic system was recorded using an optical system; the beads' movements were compared with the simulated trajectories and gave a good agreement. The model was used to predict the shortest agglomeration time for a given separation distance which is of practical benefit to users of bead based assays. The concluding part of this thesis expanded the simulation into a three dimensional model to predict the interactions among three super-paramagnetic beads within a magnetophoretic system. In order to determine the height of the magnetic beads, a Huygens-Fresnel model was implemented in Matlab which was compared with off-focused diffracted images of the beads viewed under an optical system. A good comparison was obtained by comparing the simulated three-dimensional trajectories with experimental data.

O principal objetivo do presente trabalho foi compreender a dinâmica industrial do Aglomerado Urbano-Industrial de Londrina, considerando que no processo de reestruturação produtiva, as transformações vão se manifestar também nas cidades médias. Assim como o estado do Paraná o Aglomerado Urbano-Industrial de Londrina, formado pela cidade média de Londrina e seu entorno, a partir de 1992 passou a apresentar desenvolvimento industrial importante., porém acentuando o caráter desigual deste desenvolvimento. A dinâmica industrial recente apresenta dois processos distintos: o primeiro está relacionado ao desenvolvimento de setores industriais tradicionais na região, intensivos em mão-de-obra e diretamente relacionados a empresas de pequeno e médio porte, cuja origem do capital é local/regional, além dos ramos industriais diretamente vinculados a forte base agropecuária da região. O segundo está relacionado ao processo de concentração e centralização social e espacial da reprodução do capital, vinculando-se ao processo de transferências industriais, em geral da Região Metropolitana de São Paulo. Na maioria dos casos as empresas foram beneficiadas pelos incentivos fiscais do governo estadual, por intermédio do Programa Paraná Mais-Empregos e invariavelmente receberam incentivos municipais. De fundamental importância para a dinâmica industrial do Aglomerado foi a provisão das condições gerais de produção historicamente estabelecidas, que propiciaram uma intensa integração, quer entre os municípios que o compõe, quer intensificando e modificando as relações históricas estabelecidas com o estado de São Paulo e sua região metropolitana, pólo hegemônico industrial brasileiro. Isso nos leva a reforçar a hipótese inicial de nosso trabalho de que, as condições gerais de produção historicamente estabelecidas têm um peso fundamental no desenvolvimento regional e que, a atuação do Estado na provisão dessas condições tem um alcance social e regional amplo, enquanto os programas vinculados à guerra fiscal são essencialmente concentradores, social e regionalmente. / The main objective of this work is to understand the industrial dynamics of (Aglomerado Urbano-Industrial de Londrina), considering that in the process of productive reorganization, the transformations go to also disclose themselves in the medium cities. The Aglomerado Urbano-Industrial de Londrina is formed by the medium city of Londrina and its neighborhood. As well as the state of the Paraná, the Urban-Industrial Agglomerate of Londrina, since 1992 started to present an important industrial development, however accents this unequal character. The recent industrial dynamics presents two distinct processes: the first one is related to the development of traditional industrial sectors in the region, intensive in man power and directly related the companies of small and medium size, whose origin of the capital is local/regional, beyond the industrial branches directly tied the strong farming base of the region. The second is related to the both processes: concentration and social-spatial centralization of the capital reproduction, associated to the industrial transferences process, in general of the Region Metropolitan of São Paulo. Frequently, the companies had been benefited by the tax incentives of the state government, for intermediary of the Paraná Mais-Empregos Program and certainly they had received municipal incentives. Very important for the industrial dynamics of the Agglomerated was the provision of the general conditions of production historically established, that had favored an intense integration among these cities, or intensifying and modifying the historical relations established with the state of São Paulo and its metropolitan region, the hegemonic pole industrial of Brazil. This strengthen the initial hypothesis of our work, that the general conditions of production historically established have a great importance in the regional development and the performance of the State in providing these conditions has an ample social and regional reach, while the programs related to the fiscal war are essentially social and regional concentrative.

Für nanoskalige Füllpartikel, die stark zum Agglomerieren tendieren, ist das Ultraschalldispergieren mittels Sonotrode eine geeignete Methode, um Agglomerate hinreichend gut zu entbündeln. Um dabei die optimalen Ultraschallparameter ermitteln zu können, ist es nötig, das effektive Dispergiervolumen, in welchem die Agglomerate durch Kavitation aufgebrochen werden, zu kennen. Die hier vorgestellte CAD-basierte Methode zur Berechnung des effektiven Dispergiervolumens ist dabei deutlich weniger zeitintensiv als die bisher üblicherweise verwendete analytische Methode.

No description available.

For the commercial success of polymer - multiwalled carbon nanotube (MWNT) composites the production of these materials on industrial scale by melt processing is of significant importance. The complete dispersion of primary MWNT agglomerates in a polymer melt is difficult to achieve, making it an important and challenging technological problem. Hence, it is necessary to understand the process of MWNT agglomerate dispersion in a polymer melt. Based on an intensive literature research on mechanisms and influencing factors on dispersion of other agglomerated nanostructured fillers (e.g. carbon black), the main dispersion steps were evaluated and investigated concerning the agglomerated MWNT.Consequently, systematic investigations were performed to study the effect of the melt infiltration on MWNT agglomerate dispersion and to analyse the corresponding main dispersion mechanisms, namely rupture and erosion. The states of MWNT agglomerate dispersion were assessed by quantifying the agglomerate area ratio and particle size distribution using image analysis of optical transmission micrographs. Additionally, the composite’s electrical resistivity was determined. In the prevailing study, polycarbonates (PC) varying in molecular weight were used to produce composites containing 1 wt% MWNT (Baytubes C150HP) as model systems and a discontinuous microcompounder was applied as melt mixing device. The agglomerate structure of the used MWNT material made them especially suitable for the reported investigations. The step of melt infiltration into the primary nanotube agglomerates plays a crucial role for their dispersion in the PC melt. During melt mixing when low shear rates were applied, better state of MWNT dispersion was obtained in high viscosity matrices because applied shear stresses were high. On the contrary, if high shear rates were applied, similar states of MWNT dispersion were obtained in low and high viscosity matrices although significantly lower shear stresses were applied in the low viscosity matrix as compared to the high viscosity matrix. The results indicate that if the applied shear stress values are compared, with increasing matrix viscosity the agglomerate dispersion gets worsen. This is attributed to the fact that low viscosity matrices can infiltrate relatively faster than high viscosity matrices into the agglomerate making them weaker and reducing the agglomerate strength. Thus, at sufficient shear rates MWNT agglomerates disperse relatively faster in low viscosity matrix. This illustrates a balance between the counteracting effects of viscosity on agglomerate infiltration and agglomerate dispersion. Additionally, the effect of matrix molecular weight on the size of un-dispersed MWNT agglomerates was investigated. Under similar conditions of applied shear stress, the composites based on low molecular weight matrix showed smaller sized un-dispersed primary agglomerates as compared to composites with higher molecular weight matrices. This again highlights the role of matrix infiltration as the first step of dispersion. Following the step of melt infiltration, agglomerate size gets reduced due to the dispersion mechanisms. To analyse the corresponding contributions of different dispersion mechanisms (rupture and erosion), the kinetics of MWNT agglomerate dispersion was investigated. If high mixing speeds are employed dispersion is quite fast and needs less time as compared to low mixing speed. A model is proposed to estimate the fractions of rupture and erosion mechanisms during agglomerate dispersion based on the kinetic study in the discontinuous mixer. Under the employed experimental conditions, at high mixing speeds, the dispersion was found to be governed by rupture dominant mechanism, whereas at low mixing speeds the dispersion was controlled by both mechanisms. As far as electrical resistivity is concerned, for a given content of MWNT as the state of dispersion improves, the resistivity values decrease significantly but only up to a plateau value. The composites produced using low viscosity matrices have lower resistivity values as compared to high viscosity matrices. Additionally, composites were prepared using additives, whereas the additives were found to be useful for improving filler dispersion and electrical conductivity.

For the commercial success of polymer - multiwalled carbon nanotube (MWNT) composites the production of these materials on industrial scale by melt processing is of significant importance. The complete dispersion of primary MWNT agglomerates in a polymer melt is difficult to achieve, making it an important and challenging technological problem. Hence, it is necessary to understand the process of MWNT agglomerate dispersion in a polymer melt. Based on an intensive literature research on mechanisms and influencing factors on dispersion of other agglomerated nanostructured fillers (e.g. carbon black), the main dispersion steps were evaluated and investigated concerning the agglomerated MWNT.Consequently, systematic investigations were performed to study the effect of the melt infiltration on MWNT agglomerate dispersion and to analyse the corresponding main dispersion mechanisms, namely rupture and erosion. The states of MWNT agglomerate dispersion were assessed by quantifying the agglomerate area ratio and particle size distribution using image analysis of optical transmission micrographs. Additionally, the composite’s electrical resistivity was determined. In the prevailing study, polycarbonates (PC) varying in molecular weight were used to produce composites containing 1 wt% MWNT (Baytubes C150HP) as model systems and a discontinuous microcompounder was applied as melt mixing device. The agglomerate structure of the used MWNT material made them especially suitable for the reported investigations. The step of melt infiltration into the primary nanotube agglomerates plays a crucial role for their dispersion in the PC melt. During melt mixing when low shear rates were applied, better state of MWNT dispersion was obtained in high viscosity matrices because applied shear stresses were high. On the contrary, if high shear rates were applied, similar states of MWNT dispersion were obtained in low and high viscosity matrices although significantly lower shear stresses were applied in the low viscosity matrix as compared to the high viscosity matrix. The results indicate that if the applied shear stress values are compared, with increasing matrix viscosity the agglomerate dispersion gets worsen. This is attributed to the fact that low viscosity matrices can infiltrate relatively faster than high viscosity matrices into the agglomerate making them weaker and reducing the agglomerate strength. Thus, at sufficient shear rates MWNT agglomerates disperse relatively faster in low viscosity matrix. This illustrates a balance between the counteracting effects of viscosity on agglomerate infiltration and agglomerate dispersion. Additionally, the effect of matrix molecular weight on the size of un-dispersed MWNT agglomerates was investigated. Under similar conditions of applied shear stress, the composites based on low molecular weight matrix showed smaller sized un-dispersed primary agglomerates as compared to composites with higher molecular weight matrices. This again highlights the role of matrix infiltration as the first step of dispersion. Following the step of melt infiltration, agglomerate size gets reduced due to the dispersion mechanisms. To analyse the corresponding contributions of different dispersion mechanisms (rupture and erosion), the kinetics of MWNT agglomerate dispersion was investigated. If high mixing speeds are employed dispersion is quite fast and needs less time as compared to low mixing speed. A model is proposed to estimate the fractions of rupture and erosion mechanisms during agglomerate dispersion based on the kinetic study in the discontinuous mixer. Under the employed experimental conditions, at high mixing speeds, the dispersion was found to be governed by rupture dominant mechanism, whereas at low mixing speeds the dispersion was controlled by both mechanisms. As far as electrical resistivity is concerned, for a given content of MWNT as the state of dispersion improves, the resistivity values decrease significantly but only up to a plateau value. The composites produced using low viscosity matrices have lower resistivity values as compared to high viscosity matrices. Additionally, composites were prepared using additives, whereas the additives were found to be useful for improving filler dispersion and electrical conductivity.

No description available.