The primary objective of this thesis was to develop a nucleation regime map to determine the controlling nucleation mechanism as a function of material properties and operating parameters. Two distinct regimes of nucleation were identified. The drop controlled nucleation regime occurs when nucleation conditions are ideal and one drop produces one nucleus granule and the controlling property is the droplet size. The nuclei formation kinetics are fast and the binder droplets penetrate into the powder bed pores almost immediately. In this region, the nuclei distribution reflects the drop size distribution as one drop tends to form one granule provided that (a) drops hitting the powder surface do not overlap - low spray flux Ya and (b) the drop must wet into the bed completely before bed mixing brings it into contact with another partially absorbed drop on the bed surface - low penetration time tp. If either criterion is not met, powder mixing characteristics will dominate. In the mechanical dispersion regime, the viscous or poorly wetting binder is slow to flow through the powder pores and form nuclei and good mixing is required for binder dispersion. The kinetics of nuclei formation were characterised using a simple drop penetration time test. A single drop of binder fluid was placed on a loosely packed powder bed and the time taken for the fluid to penetrate completely was measured for a range of powder and binder combinations. Loosely packed powder beds contain large macrovoids which are included in the existing Kozeny approach to estimating pore size. However it was found that these pores do not participate in liquid flow. A new two phase model was proposed where the total volume of the macrovoids was assumed to be the difference between the bed porosity and the tap porosity. A new parameter, the effective porosity eeff, was defined as the tap porosity multiplied by the estimated fraction of pores that terminate at a macrovoid and are effectively blocked pores. The pore sizes and drop penetration times were recalculated using the effective porosity and the predicted tp values were generally within an order of magnitude of the experimental results for all powders. The drop penetration time is reduced by small drops, low viscosity fluids, porous powders (but without macrovoids), large powder pores, high surface tension, low contact angle and pre-wetting of the powder bed. A new dimensionless group, the dimensionless spray flux Ya was defined to characterise the three most important operating variables in binder dispersion: liquid binder flowrate, drop size and powder flux through the spray zone. At low Ya, the majority of drops land on the powder sufficiently well separated to allow ideal “drop controlled” nucleation where one drop forms one granule. As Ya increases, the probability of drop footprints overlapping to give larger agglomerate nuclei increases. Monte-Carlo simulations were performed to validate the spray flux theory. The proportion of nuclei formed from single drops falls exponentially as Ya increases and to remain in the drop controlled regime Ya must be kept below 0.1. Analytical solutions based on the Poisson distribution for the fraction of single drop nuclei as a function of Ya were an excellent match with the Monte-Carlo data. Further validation experiments in carefully designed ex-granulator experiments and in an industrial granulator were performed. The results matched the theoretical solutions and demonstrated the ability of Ya to describe the nucleation zone in a real granulator. The proposed nucleation regime map demonstrated the interaction between drop penetration time and spray flux in nucleation. At short penetration times, such as the water and lactose system, decreasing Ya causes a shift towards the drop controlled regime and a narrower nuclei distribution. When penetration time is long, the nuclei size distribution is always larger and broader. Granulation may still be successful if the mechanical dispersion forces are able to break up the binder clumps and distribute the binder through the powder. The nucleation regime map should prove to be a useful tool for maintaining effective liquid distribution during scale-up as well as a useful trouble-shooting tool. It allows the dominant mechanism controlling the nucleation process to be easily identified using relatively simple parameters and a rational approach can then be used to control the properties of the nuclei.

This thesis covers a systematic examination of wet granulation in a twin screw extruder. Granulation of the excipient, (alpha)-lactose monohydrate, was done with the aid of PVP in an aqueous solution which acted as a binding agent. The influences on agglomeration by the following processing parameters were studied: screw elements design, screw rotational speed, binding solution concentration, and binder addition method. Qualitative efforts had also been made in modeling the process to gain valuable insight into how the elements affected agglomeration and granule rupture. A commercial software package PFC^2D, based on the Discrete Element Method (DEM), was used to simulate the dynamic behavior of the screw elements in the barrel. Within the optimal range of 7.5 -10wt% binder concentration, all the screw profiles were studied for their capacity to produce desirable granules suited to solid oral dosage form production. By increasing the rotational speed from 30 RPM to 80 RPM, the granules size of the conveying, discharging and chopping elements decreased whereas this operating parameter had little effect on granule size within kneading blocks. The nominal particle size produced by a screw element increased from 300(mu)m to 1mm when dispersive mixing was its dominant purpose (i.e. the kneading block), thereby meeting our criteria for a suitable granule in tab letting. Similar size development of the granules was not found with the other conveying or distributive mixing elements. In regards to particle shape, the kneading blocks produced elongated shape granules while other elements tested in this study produced smaller, more spherical agglomerates. Either shape was found effective in tabletting. Wet granulation was not feasible with more extreme concentrations of the aqueous binder (i.e. 5 wt% or 12 wt%) in this project, and the hand pre-blend method was the only approach found suitable for metering this additive into the system while maintaining steady feeding rates and output. / Thesis / Master of Applied Science (MASc)

Thesis (PhD (Food Science))--University of Stellenbosch, 2009. / In the absence of anaerobic granules, anaerobically digested sewage sludge is frequently used to seed industrial upflow anaerobic sludge blanket (UASB) reactors. Because of its flocculent nature, start-up with digested sludge instead of granular sludge proceeds much slower and presents various operational problems. Any manner in which the granulation of digested sludge can be enhanced would benefit UASB reactor start-up and application in developing countries such as South Africa. The main objective of this dissertation was to improve granulation and reduce UASB reactor start-up by using pre-treated digested sludge as seed. The sludge was pre-treated based on the batch granulation-enhancement model of Britz et al. (2002). The main aim of the model was to improve extracellular polymer (ECP) production of lactate-utilising populations by applying short-term controlled organic overloading in a mechanically agitated environment. The batch granulation-enhancement (pre-treatment) process was applied to an ECP-producing digester strain, Propionibacterium jensenii S1. Non-methanogenic aggregates were formed when batch units were incubated on a roller-table instead of a linear-shake platform. Larger, more stable aggregates were obtained in the presence of apricot effluent medium. Preliminary batch granulation-enhancement studies confirmed that using the roller-table as mixing system had a positive influence on batch granulation-enhancement. The roller-table showed the most potential for handling larger volumes in comparison to a linear-shake waterbath and linear-shake platform. The addition of 450 mg.L-1 Fe2+ at the start of the study also influenced aggregate numbers positively. These studies revealed that pre-treatment results varied depending on the seed sludge source. A denaturing gradient gel electrophoresis (DGGE) method was applied for the detection of Archaea in digested sludges and UASB granules. In addition, a methanogenic marker containing methanogens important to the granulation process was constructed to aid identification. The positive influence of DMSO and “touchdown” PCR on the elimination of artifactual double bands in DGGE fingerprints were also demonstrated. Results revealed that only one of the four digested sludges tested contained Methanosaeta concilii (critical to granular nuclei formation) while it was present in all the UASB granules regardless of substrate type. Four digested sludges were obtained from stable secondary digesters. DGGE indicated the presence of M. concilii in all sludges. The Athlone 4Sb-sludge was the only sludge which exhibited measurable methanogenic activity during substrate dependent activity testing. The ST-sludge showed the highest increase in volatile suspended solids (VSS) particles ≥0.25 mm2. Laboratory-scale UASB reactor start-up was done with both sludges and start-up proceeded better in the Athlone 4Sb-reactor. Athlone 4Sb-sludge batches were pre-treated in a rolling-batch reactor in the presence of either lactate or sucrose and used to seed lab-scale UASB reactors B (sucrose seed) and C (lactate seed). Start-up efficiencies were compared to a control (Reactor A). Overall Reactor B was more efficient that the control. At the end of the study the Reactor B sludge had a higher methanogenic activity than the control reactor. It also had the highest increase in VSS ≥1.0 mm2. Pre-treatment of digested sludge in the presence of sucrose, therefore, aided granulation and reduced UASB reactor start-up time.

Yes / Several contributions in answering granulation challenges including the use of computer simulation and well thought out experimental analyses are being researched. Using a twin screw granulator (TSG) by design of experiments (DoE), comparisons on 1) equipment similarities i.e., continuous and 2) shear forces, are made to previous literature on continuous equipment and a Cyclomix. This study proposes that equipment specific DoE, better explains the contribution of parameters than investigating an identified parameter from the experimental findings from a specific equipment. Granule strength and structure are presented together with the contribution of process parameters, speed, temperature, and binder content. Seeded structures are present in all but the Extrudomix. Longer residence times within the Cyclomix facilitates seeded structures. Granule crushing strengths are higher in TSG than all other continuous equipment. Optimum condition for the formation of stronger granules with least variation is around 65.4 °C. / The authors would like to acknowledge the support from the CCIP grant (Collaboration, Capacity and IP Development) fund from the University of Bradford for ordering cunsumables and equipment.

Stellar model atmospheres are a fundamental tool for our understanding of stars. Because the chemical composition of stars cannot be measured directly, the inferred stellar parameters are model dependent. In recent years great progress has been made in the modeling of stellar atmospheres, allowing the relaxation of simplifying assumptions made in previous models. The use of new 3D model atmospheres to infer the solar chemical composition has resulted in a decrease of the solar metallicity. This result has caused some controversy and is being challenged. The main aim of this thesis is to ascertain if the new models of stellar atmospheres are realistic and can be trusted to derive the chemical composition of stars in general, and the Sun in particular. Other objectives also include the study of line formation in the Sun at high spatial resolution, and possible implications in the modeling. The Sun is the ideal test-bench for detailed analyses of stellar atmospheres. With the solar surface being resolved in great detail and at different viewing angles, a wealth of information can be gathered that allows for very robust tests of atmosphere models. The testing detailed here addresses several fronts. On one hand, the model�s temperature structure is directly tested with the classical tests of continuum centre-to-limb variations and absolute fluxes. On the other hand, the line formation is tested at different viewing angles and high spatial and spectral resolution. Here the main focus is on oxygen lines, as oxygen has an important contribution for the total solar metallicity. However, other lines are also tested. High quality data were specifically obtained for these line formation tests, using the Swedish 1-m Solar Telescope. For the temperature structure tests we find a surprisingly good agreement between the 3D model and the observations, surpassing even semi-empirical models. The solar 1D non-LTE models have a very similar behaviour to 1D LTE models, confirming that LTE is a good approximation in the solar photosphere. The 3D theoretical model performs consistently better than its 1D counterparts. The oxygen line formation tests are carried out in great detail, with a careful wavelength calibration, revised atomic data, and allowing for departures from LTE. Again we find a reassuring agreement between the 3D model predictions and the observations, both for the centre-to-limb variation of the lines and the line formation at high spatial resolution. The observations at different viewing angles also allowed the empirical determination of the role of hydrogen collisions with oxygen, important when deriving the oxygen abundance. The tests undertaken here show that the 3D model atmospheres are indeed very realistic. Their predicted temperature structure and velocity fields compare very favourably with observations of the Sun. Together with previous tests, this indicates they can be relied upon to derive the chemical composition of the Sun and similar late-type stars.

Orientador: Florencia Cecilia Menegalli / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia de Alimentos / Made available in DSpace on 2018-08-08T12:47:34Z (GMT). No. of bitstreams: 1 Souza_DiogoOtaviodeCastro_M.pdf: 3487694 bytes, checksum: 3590d698286e2f74b76a3cd472ba2d1f (MD5) Previous issue date: 2007 / Resumo: O objetivo do presente trabalho é a caracterização morfológica das partículas de suco de goiaba em pó, granuladas em um leito fluidizado, utilizando água como líquido ligante, além de verificar aplicação do modelo de balanço populacional na modelagem do crescimento destas partículas, ao longo da granulação. Este estudo abrange a determinação do número mínimo de partículas a serem analisadas para determinação da distribuição percentual, dos fatores de forma e do número de partículas por unidade de massa de material; abrange o ajuste dos parâmetros do modelo e a seleção do melhor mecanismo que descreve o processo. Para a determinação do número mínimo de partículas a serem avaliadas, foi utilizada, para a distribuição percentual, o teste de Qui-Quadrado; para os fatores de forma, foi utilizada variação dos desvios-relativos; e para o número de partículas por unidade de massa de material, um delineamento inteiramente casualizado. Foi determinado que o número de 550 partículas seria o mínimo necessário para a determinação da distribuição percentual; 1100 para determinação dos fatores de forma e 450 para determinação do número de partículas por unidade de massa do material. A Compacidade foi o único fator de forma capaz de descrever bem a morfologia das partículas, dentro de cada intervalo de tamanho. Nenhum fator de forma foi capaz de descrever a massa total de grânulos como um todo, através da média. O modelo de balanço populacional foi resolvido através do método de segregação. O modelo se mostrou capaz de descrever as alterações ocorridas nas distribuições de tamanho das partículas ao longo do processo de granulação, mostrando que a aglomeração de partículas é proporcional à diferença de tamanho entre as partículas envolvidas nas colisões / Abstract: The aim of this work is the morphological characterization of the particles of guava¿s juice powder, granulated in a fluidized bed, using water as binder, beyond the study of the population balance model application in the modeling of the growth of the particles along the granulation. This study discusses the determination of the minimal number of particles to be analyzed for the percentual distribution estimation, the form factors estimation and number of particles per unity of material mass estimation; it also shows the adjust of model parameters and the selection of the best mechanism to describe the process. For the determination of the minimal number of particles to be evaluated, the Qui-Square test was performed, to the percentual distribution, the variation of the relative-errors to the form factors, and an experimental design completely randomized for the number of particles per unity of material mass. It was determinated that the number of 550 particles would be the minimal necessary to determination of the percentual distributions, 1100 to determination of the form factors and 450 to determination of number of the particles per material mass. The Compacity was the only form factor capable to describe well the particle¿s morphology, inside each size interval. None of the form factors were able to describe the entire mass of granules through the media. The discretized method was used to the resolution of the population balance model. The model was able to describe the size distribution alterations occurring along the granulation process, showing that the agglomeration is proportional to the size difference between the colliding particles / Mestrado / Mestre em Engenharia de Alimentos

Orientador: Sandra Cristina dos Santos Rocha / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Quimica / Made available in DSpace on 2018-08-14T18:51:32Z (GMT). No. of bitstreams: 1 Cunha_RobertoLuisGomesda_D.pdf: 5308810 bytes, checksum: 23031bb71d61ff09f951cb9f06de28e9 (MD5) Previous issue date: 2009 / Resumo: Atualmente a granulação de partículas finas encontra suas principais aplicações nas indústrias de alimentos, química, agrícola e especialmente farmacêutica e tem sido cada vez mais objeto de pesquisa, com finalidades de melhorar aspecto, perdas por manuseio, compactação e transporte. A granulação consiste em aumentar o tamanho (volume) da partícula, através da incorporação de materiais ativos e/ou inertes. Os primeiros trabalhos foram desenvolvidos na década de 50, utilizando areia em granuladores cilíndricos. Desde então, uma grande quantidade de trabalhos vem sendo realizado com a utilização de equipamentos tais como misturadores, tambores rotativos e leito fluidizado. A utilização do leito fluidizado modificado tem como objetivo melhorar o processo, otimizando a técnica e tornando-a economicamente viável. Neste trabalho, estudou-se a granulação de pós em leito fluidizado com tubo interno, utilizando uma solução aquosa com concentração de 35 % de maltodextrina como agente granulante, e efetuou-se a caracterização física dos grânulos formados e de pastilhas obtidas a partir desses grânulos. O material empregado como excipiente foi a celulose microcristalina com diâmetros que variam de partículas menores que 125 mm a 420 mm. O leito consiste de uma coluna em acrílico de 60 cm de altura e 14 cm de diâmetro interno. Um tubo com 14 cm de altura e 7 cm de diâmetro interno foi adaptado no centro do leito, preso ao topo por uma haste retrátil, facilitando o seu deslocamento e fixação em relação à placa distribuidora. A utilização do leito fluidizado com um tubo interno facilitou a condução do processo em relação ao leito fluidizado convencional, pois foi observada uma melhora na fluidodinâmica durante o processo. A solução granulante foi aspergida sobre o leito de partículas, através de um bico atomizador de duplo-fluido. As variáveis operacionais analisadas foram: vazão da solução granulante (Ws), com valores de 4, 7 e 10 ml/min no display da bomba, pressão de atomização (Pat) de 68,95, 103,42 e 137,89 kPa, temperatura do ar de fluidização (T) de 60, 70 e 80 °C, e altura do tubo interno (ht) em relação à placa distribuidora de 4, 6 e 8 cm. O trabalho experimental foi realizado com auxílio de um planejamento fatorial completo 24 para avaliar a influência das condições operacionais nas variáveis de resposta: crescimento das partículas, índice de torrões e índice de Carr. O planejamento experimental forneceu modelos estatisticamente significativos para as respostas crescimento dos grânulos e índice de Carr. O crescimento dos grânulos foi favorecido pelo aumento da vazão da solução granulante associada à baixa pressão de atomização. Os grânulos apresentaram compressibilidade de 2,44 a 16,9 % e ângulo de repouso de 37,6±1,77° a 42±1,49°, indicando ótima fluidez e coesividade intermediária. Além disso, analisou-se também a compressão e compactação dos grânulos. Pastilhas foram confeccionadas aplicando 1,5 ton. aos grânulos em prensa hidráulica e submetidas à análise de resistência. Os resultados mostraram-se satisfatórios, estando os valores obtidos de acordo com os especificados na literatura. / Abstract: Currently the fine particle granulation finds its main applications in the food chemical, agricultural and especially pharmaceutical industries, and has been attracting more and more interest with purposes to improve aspect, losses, compacting and transport. The granulation consists of increasing the size of the particle, through the incorporation of active and/or inert materials. The first works had been developed in the decade of 50. Since then, a great amount of works comes, being carried through with the use of different equipments such as mixers, rotary drums and fluidized beds. Modified fluidized bed granulators had been proposed with the objective of improving the process. The use of modified fluidized beds optimizes the technique and turns it economically viable. The objective of this work was the study of granulation of powders in a fluidized bed modified with the inclusion of a draft-tube, using an aqueous solution of maltodextrin (35% in weight), and the physical characterization of the granules and tablets compressed from these granules. The granulation operation was carried out by top spraying the granulating liquid on the fluidized bed of particles using a double-fluid nozzle. Microcrystalline cellulose particles (mean diameter from below 125µm to 420µm) were chosen as the raw material to be granulated. The fluidized bed consisted of a cylindrical column constructed in Plexiglass with 0.14 m of internal diameter and 0.6 m of height. A draft-tube with 0.14 m of height and 0.07 m of internal diameter was adapted in the center. The experimental work was carried through with the aid of a 24 experimental design with three central points, resulting in 19 runs to evaluate the effects of the operational variables on the granulation process. The independent variables studied were the flow rate of granulating solution (Ws); the air pressure of atomization (Pat); the inlet air temperature (T) and the height of the drafttube (ht). The values of the variables at the lower, central and upper levels were fixed in 4, 7 and 10 mL/min for Ws; 68.95, 103.42 and 137.89 kPa for Pat; 60, 70 and 80°C for T and 4, 6 and 8 cm for ht. The responses analyzed were the particles growth, agglomeration index, angle of repose and Carr index. The growth of granules was favored by the increase of the flow rate of the granulating solution associated to a low pressure of atomization. The granules showed compressibility in the range of 2.44 to 16.9% and angle of repose from 37.6±1.8° to 42.0±1.5°, indicating free flow ability and intermediate cohesiveness. Also, the granules compression and compaction was analyzed. Tablets were done by applying 1.5 tons to the granules using a hydraulic press and then submitted to resistance analyses. The results obtained were satisfactory as the values were in the range specified in the literature for pharmaceutical tablets. / Doutorado / Engenharia de Processos / Doutor em Engenharia Química

Aspen Pharmacare, one of the leading pharmaceutical manufacturers in South Africa has embarked on a programme of improving the production processes currently employed at their Port Elizabeth site. With the introduction of new technology at the site and the move towards globalization, it became imperative that Aspen remain competitive in the market. The product of interest in this research, Degoran Plus tablets, is one of the company’s leading brand sellers. Upon investigation, it became apparent that this product created opportunity for process improvement using the new technology. The manufacture of Degoran Plus entails granulation, compression and coating of the product. Most opportunity for improvement was possible in the granulation stage because of the laborious nature of the present process. Degoran Plus tablets had a history of analytical failures, especially with regard to the dissolution rate of the final product, as well as other quality related issues. The product was not considered to be a “through-runner”, which resulted in bad production output, due to continual repeats of not only analysis but also reworks in production. A strategic decision was taken to manufacture Degoran Plus using the Collette Gral granulator as the equipment offered superior mixing capability when compared to the Bear planetary granulator. It was assumed that the granulation process would result in more uniform distribution of the actives. Upon producing a better granule, a final product of superior quality would be attained. The validation protocol stipulates that three samples be taken and tested from the powder mix. Nine samples taken from granulated bulk are treated in the same manner. The validation protocol further stipulates that the first three batches manufactured utilise the new process, and tested according to the protocol. The results obtained from the analysis are evaluated statistically and a conclusion and recommendation were derived based on the evaluation.

The enigma of anaerobic sludge granulations is still exciting the minds of both experimental scientists and modeling experts. A unique combination of mechanical, physiochemical and biological forces influence granulation during processes of anaerobic digestion. However, knowledge of potential driving forces of granulation has not been transformed into a comprehensive model of anaerobic granulation. In this computational experiment, we address the role physiochemical and biological processes play in granulation and provide a literature-validated working model of anaerobic granule de novo formation. The model developed in a cDynoMiCs simulation environment successfully demonstrated a de novo granulation in a glucose fed system. The simulated granules exhibit experimental observations of radial stratification: A central dead core surrounded by methanogens then encased in acidogens. Practical applications of the granulation model was assessed on the anaerobic digestion of low-strength wastewater by measuring the changes in methane yield as model parameters were systematically swept. This model will be expanded in the future to investigate the influence of mechanical forces on the de novo granulation and the application of a model to anaerobic digestion of a complex protein-carbohydrate rich feedstock.

No description available.