Stuttgart, Univ., Diss., 2001.

Recent publications on ethanol production and purification shows optimized energy and water consumptions as low as 22,000 Btu/gal ethanol and 1.54 gal water/gal ethanol respectively using multieffect distillation. Karuppiah, et al use column rating and mathematical optimization methods and shortcut design models to design evaluate and optimize the energy and water consumption. In this work, we compare shortcut design and rigorous simulation models for an ethanol purification distillation system, and we show that distillation systems based on shortcut design underestimate the true energy and water consumption of the distillation system. We then use ASPEN Plus, to design a multieffect distillation system and an extractive distillation system using rigorous simulation and compare the two for energy and water consumptions. We show that the extractive distillation system has lower steam and cooling water consumptions and consequently lower energy and water consumptions than multieffect distillation in corn-to-ethanol production and purification. We also show that the extractive distillation system is cheaper than the multieffect distillation system on a cost per gal ethanol basis. This work gives an energy consumption of 29987 Btu/gal ethanol and water consumptions 2.82 gal/gal ethanol for the multieffect distillation system at a manufacturing cost of $3.03/gal ethanol. For the extractive distillation system, we calculate an energy consumption of 28199 Btu/gal ethanol and a water consumption of 2.79 gal/gal ethanol at a manufacturing cost of $2.88/gal ethanol. / Master of Science

Orientador: Maria Regina Wolf Maciel / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-18T13:50:07Z (GMT). No. of bitstreams: 1 JaimesFigueroa_JaiverEfren_M.pdf: 4284661 bytes, checksum: 5b7383def90a5143160bc892b406c75c (MD5) Previous issue date: 2011 / Resumo: A produção de etanol a partir da cana de açúcar é uma tecnologia dominada completamente pelo Brasil porém, encontra-se na etapa de intensificação, otimização e inovação. O etanol pode ser produzido como hidratado ou anidro, sendo necessário, para produção deste último, um processo posterior de desidratação. Existem inúmeros processos de desidratação, dentre dos quais a destilação extrativa é um dos mais simples de realizar. A destilação extrativa usa um solvente para modificar o equilíbrio líquido-vapor, permitindo quebrar o azeótropo etanol/água que impede que a desidratação seja feita por destilação convencional. O solvente de extração é de grande importância, dele depende a facilidade com que vai ser feita a separação, a quantidade a ser utilizada e o requerimento energético do processo. Nesse contexto, aparecem os líquidos iônicos, que são apresentados como ótimos solventes potenciais de extração; um líquido iônico (LI) é um sal composto por um cátion orgânico com pelo menos uma carga deslocada e um ânion inorgânico; sua estrutura evita que se forme uma rede cristalina estável, resultando em solventes líquidos altamente iônicos com temperaturas de fusão inferiores a 100 °C e com insignificante pressão de vapor. Os LI são principalmente usados em substituição aos solventes convencionais, podendo ser uma alternativa para diminuir a poluição ambiental, evitando a emissão de componentes orgânicos voláteis ao meio ambiente. Com a justificativa anterior, o objetivo desta dissertação foi analisar e otimizar o processo de obtenção de etanol anidro a partir da mistura etanol + água de composição pré-azeotrópica, empregando líquidos iônicos (LI), visando avaliar seu potencial; os LI estudados foram: 1-butil-3-metilimidazólio cloreto, 1-butil-3-metilimidazólio metilsulfato, 1-butil-3-metilimidazólio acetato, 1-butil-3-metilimidazólio tetrafluoroborato, 1-butil-3-metilimidazólio dicianamida, 1-etil-3-metilimidazólio cloreto, 1-etil-3-metilimidazólio tetrafluoroborato, 1-hexil-3-metilimidazólio cloreto. Neste trabalho foi encontrado o requerimento energético e a quantidade de LI a ser empregado para obter os valores máximos de pureza e porcentagens de recuperação de etanol e água. A influência das condições de operação e desenho utilizadas, tais como fração de etanol na alimentação, relação LI:alimentação, temperatura da alimentação e do LI de reposição, quantidade de estágios, relação molar de refluxo, estágio de alimentação e vazão molar de destilado da coluna de recuperação de etanol e de purificação de LI, foram analisadas empregando o simulador comercial Aspen Plus e, otimizadas empregando a técnica de delineamento de experimentos. Todos os LI estudados apresentaram capacidade de desidratar o etanol, elevando sua concentração de pré até pós-azeotropia, obtendo-se pureza de etanol maiores que 0,995 em massa. Além disso, dependendo do LI utilizado, o processo atinge porcentagens de recuperação de etanol e água, em média, de 98% e 74%, respectivamente. Na definição do modelo para o coeficiente de atividade do equilíbrio ternário líquido vapor da mistura etanol + água + LI foram testados o NRTL e UNIQUAC, chegando-se à conclusão de que o equilíbrio representado pelo modelo de NRTL é o mais adequado / Abstract: The production of ethanol from sugar cane is a technology led and dominated by Brazil. However, it is still in a stage of optimization and innovation. Ethanol can be produced in a hydrated or dehydrated state, but the latter requires an additional process to the conventional distillation. There are numerous dehydration processes that can be implemented, but the extractive distillation is one of the most simple. Extractive distillation uses a solvent that modifies the liquid-vapor equilibrium and eliminates the presence of the ethanol-water azeotrope that prevents the use of conventional distillation for the dehydration process. The solvent for the extraction is of great importance since it dictates the degree of separation and the energy requirements for the process. In this context, ionic liquids are considered since they have been presented as excellent solvents for extraction. An ionic liquid (IL) is a salt formed by an organic cation with at least one delocalized charge, and an inorganic anion. The structure of the ionic liquids prevents the formation of a stable crystalline net, resulting in highly ionic liquid solvents that have melting points below 100 ºC and negligible vapor pressures. With those characteristics, ionic liquids can be a replacement for conventional solvents offering alternatives for the decrease of the environmental impact by preventing the emissions of volatile compounds to the environment. With the previous justification, the objective of this master dissertation was to analyze and optimize the process of obtaining anhydrous ethanol from a mixture ethanol + water with pre-azeotropic composition by using ionic liquids; and also to evaluate their performance in this application to evaluate its potential. Ionic liquids were studied: 1-butyl-3-methylimidazolium chloride, 1-butyl-3-methylimidazolium methylsulfate, 1-butyl-3-methylimidazolium acetate, 1-butyl-3-methylimidazolium tetrafluoroborate, 1-butyl-3-methylimidazolium dicyanamide, 1-ethyl-3-methylimidazolium chloride, 1-ethyl-3-methylimidazolium tetrafluoroborate, 1-hexil-3-methylimidazolium chloride. In this work, the minimum energy requirement and the amount of ionic liquid needed to obtain maximum ethanol purity and maximum recovery of ethanol and water exiting the process were found. The influence of the design and operation conditions used, such as the ethanol composition in the feed, the IL/feed ratio, the temperature of the feed and the IL, the number of plates, the reflux molar ratio and the distilled flux in the columns of purification of ethanol and recovery of ionic liquids were studied using the commercial simulator ASPEN PLUS, and optimized by utilization of the design of experiments (DOE) technique. All the ionic liquids used were able to dehydrate the ethanol, increasing its concentration from pre to post azeotrope, generating ethanol with purity above 0.995 in mass. In addition to that, depending on the ionic liquid used, the process reached average water and ethanol recoveries of 98% and 74% respectively. In the definition of the model for the activity coefficient in the ternary vapor-liquid equilibrium of the ethanol-water-IL mixtures, the models NRTL and UNIQUAC were studied concluding that the NRTL model was the most adequate / Mestrado / Desenvolvimento de Processos Químicos / Mestre em Engenharia Química

Repellents play a key role in preventing mosquito-borne diseases such as malaria by reducing human-vector contact. The general mechanism of action relies on providing a repelling vapour around the applied area on the skin. Thus, the proper evaporation rate and consistency of the composition of the released vapour are factors determining the performance of repellent formulations. The formulation should evaporate fast enough to provide a sufficient level of repellence during its life time. However, if evaporation proceeds too fast, then it will be depleted rapidly so that activity is lost within a short period of time, which makes the repellent inefficient. Several controlled-release approaches have been developed to improve both the protection time and level. However, these techniques have inherent drawbacks from the industrial point of view. Moreover, these techniques mostly focus only on reducing the release rate, while the consistency of the vapour composition has not been addressed. In the present study, a novel approach towards controlling the evaporation behaviour of repellents is proposed. It is based on engineering the molecular interactions in order to design negative pseudo-azeotrope formulations. Negative pseudo-azeotrope mixtures are less volatile than the pure parent components and they do not undergo separation during evaporation. The feasibility of the idea was investigated by studying the molecular structure of generally available repellents. Among known molecular interactions, hydrogen bonding has the most likely impact on the formation of azeotropes and in particular pseudo-azeotropes. Thus, established repellents were classified based on their chemical structures and their capability to take part in hydrogen bonding. Next, a simple spectroscopic method for anticipating pseudoazeotropes formation was developed. Binary compositions of nonanoic acid and ethyl butylacetylaminopropionate (IR3535) showed a potential for forming pseudo-azeotrope mixtures. Hence R3535 and nonanoic acid were selected as model compounds to test the hypothesis. An experimental technique to confirm pseudo-azeotrope formation and to locate the composition of the probable pseudo-azeotrope point was required. To this end, an oven test was designed. The temporal mass loss, under an isothermal program, of a series of evaporating mixtures was measured. Simultaneously, the Fourier transform infrared (FTIR) spectra of the liquid remaining was recorded. Inverse analysis techniques were used to determine the composition of remaining liquid mixtures from the recorded FTIR spectra. The oven tests revealed that, as vaporisation progressed, the composition of the liquid remaining and the emitted vapour converged to a fixed IR3535 content of ca. 75 mol%. Mixtures close to this composition also featured the lowest volatility. Oven test also showed that the composition of the liquid mixtures diverged from the fixed IR3535 content of ca. 10 mol%. Mixtures close to this composition featured the highest volatility. These observations showed that IR3535 and nonanoic acid forms two pseudo-azeotrope compositions, i.e. a negative pseudo-azeotrope at an IR3535 content of ca. 75 mol%, and a positive pseudo-azeotrope at IR3535 content of ca. 10 mol%. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were applied to check these results. TGA confirmed that the negative pseudo-azeotrope mixture is less volatile while the positive pseudo-azeotrope is more volatile than the parent compounds. The DSC results revealed that in comparison with the pure compounds, negative pseudo-azeotrope had a lower boiling point onset while the positive pseudo-azeotrope had a higher boiling point. Although negative pseudo-azeotrope repellent formulations have the desired lower constant release rate, their repellent activity needed to be tested. This is due to the fact that mixing the ingredients to formulate a negative pseudo-azeotrope results in interactions among the components. As a consequence, the inherent repellence effect of the compounds might have been impaired in the mixture. The modified arm-in-cage test was used to test the repellence of the controlled-release repellent formulation i.e. the negative pseudo-azeotrope of the IR3535 + nonanoic acid system. Results showed that the mixture featured improved performance with respect to both repellence efficacy and persistence. Moreover, the negative pseudo-azeotrope also exhibited a knock down effect, even resulting in mortality of most of the test mosquitoes. The presence of two pseudo-azeotrope points at different composition in the IR3535 + nonanoic acid system is a rare occurrence, analogous to double azeotropy. Thus, molecular simulation techniques were used to explore the nature of system and the interactions responsible for this unique behaviour. Gibbs-Monte Carlo simulation results suggest that variations in the sizes of the molecular clusters present in the liquid at various compositions might be responsible. They revealed that IR3535 and nonanoic acid in neat form are both highly structured liquids. The break-down in the structure of IR3535 at high concentrations of the acid may be the origin of increased evaporation rate and formation of the positive pseudo-azeotrope. On the other hand, negative pseudo-azeotrope may be resulted from formation of bulkier clusters at the ration of 3:1 (IR3535: nonanoic acid). / Thesis (PhD (Chemical Technology))--University of Pretoria, 2016. / English / PhD (Chemical Technology) / Unrestricted

Liquid mixture is everywhere in the chemical industry and widely studied by researchers. An accurate prediction of its physicochemical property is of vital importance in developing efficient process optimization. However, measurements from experiment are usually time consuming and inefficient. Furthermore, clear understanding of many of fundamental physicochemical phenomena hasn't been obtained, which restricts the development of novel products. Molecular simulation techniques have become an impressive tool to deal with these challenges during past decades. This thesis mainly applied molecular simulation to predict the physicochemical properties of industrially relevant mixtures and investigate the molecular mechanism behind observed phenomena. Among various properties, cohesive energy is the central focus, which reveals intermolecular interactions between molecules of different types. Mixture systems of two different areas of application were studied. The first is amorphous polymer-plasticizer mixtures, which, with varying composition, correspond to plastic products of different grades for application in different areas. The most important class of plasticizers are phthalate diesters, in which di (2-ethylhexyl) phthalate (DEHP) is the most frequently used compound. However, phthalates are prone to migration loss from the host poly(vinyl chloride) (PVC), which results in the contamination of surrounding environment, gradual deterioration of plastics performance, and potential harm to human health. It has thus prompted tightening governmental regulation on their usage. With this background, we aim to address three challenges: (I) model plasticized PVC to predict its physicochemical property, (II) obtain molecular insight into plasticization and plasticizer diffusion pattern inside PVC, (III) correlate plasticizer performance -- compatibility, efficacy, and mobility -- with its molecular structure. Cohesive energy plays a central role especially in understanding plasiticzer compatibility and migration tendency. Our modeling and simulation protocol is firstly tested on phthalates, where the simulated plasticization efficacy and thermodynamic compatibility with the host polymer agree well with all known experimental observations. Furthermore, through simulation of plasticizer diffusion pattern, we found relaxation of the alkyl side chains is a key factor in plasticizer migration. Next, we expand our simulation to a wider group of plasticizers including adipates, trimellitates, and citrates. The computed mixing enthalpy and Young's modulus again show an excellent agreement with available experimental data. Dependance of plasticizer performance on seven molecular design parameters are evaluated. The obtained relationship clearly tells us decreasing leg length or increasing branching on the leg will raise plasticizer compatibility with PVC, changing the torso group from benzene ring to alkane chain will highly improve plasticizer efficacy, and attaching three legs on the torso will decrease plasticizer mobility. As a side outcome, we also report a nontrivial chain-length dependence of the cohesive energy and solubility parameter of long-chain polymers, which is an important consideration in the calculation of these quantities using molecular simulation. The second area is azeotropes, the separation of which in chemical processes is usually very difficult due to the same composition in vapor and liquid phases at the azeotropic point. So far, a fundamental understanding of azeotrope formation is still missing. In this thesis, we aim to address two fundamental questions: (I) the mechanism for ethanol/benzene azeotrope formation, (II) classification of different polar-polar positive azeotropes. First, Gibbs ensemble Monte Carlo (GEMC) simulation is performed to predict the vapor-liquid equilibrium (VLE) phase diagram of ethanol/benzene, including an azeotrope point. The results match well with experiments. Free energy and cohesive energy profiles analyses are then performed. From a thorough liquid structure analysis, we conclude a three-stage mechanism for azeotrope formation: 1) formation of small ethanol clusters at low composition, 2) microscopic phase separation between ethanol and benzene, 3) isolation of benzene. This approach is then extended to four additional polar-polar mixtures (ethyl acetate/methanol, ethyl acetate/ethanol, ethanol/water, and 1-propanol/water) to obtain their VLE diagrams, which again match well with experiments. Free energy and cohesive energy analyses indicate that there are two types of mechanisms, a three-stage mechanism with weak cross-interactions (for the first two mixtures) and a three-stage mechanism with strong cross-interactions (for the last two mixtures). So far, our analyses on mixture liquid micro-structure can partially prove the existence and classification of those mechanisms. Overall, the successful prediction in physicochemical properties of two liquid mixtures with very different molecular scales proves the robustness of our study strategy, which could be used to study any liquid mixtures and understand their related physicochemical phenomena. / Thesis / Doctor of Philosophy (PhD)

Généralement les modèles QSPR ne sont utilisés que pour prédire des propriétés des corps purs. Dans cette thèse nous avons développé une approche QSPR permettant de prédire des propriétés non additives de mélanges binaires, plus précisément leur caractère azéotropique/zéotropique. Pour parvenir à ce résultat, plusieurs types de modèles quantitatifs et qualitatifs ont été développés. L’approche est originale pour deux raisons. Premièrement, peu de travaux de recherche ont été publiés sur des mélanges dont les propriétés sont non-additives. Deuxièmement, plusieurs nouveaux aspects méthodologiques ont été introduits dans ce travail. Tout d'abord des descripteurs "spéciaux", capables de décrire des mélanges ont été proposés. De plus, un protocole robuste d'obtention et de validation des modèles a été utilisé, et un domaine d'applicabilité des modèles fiable a été proposé. La méthodologie développée pendant cette thèse démontre la fiabilité d'un nouveau concept – les modèles QSPR pour les mélanges. Elle est comparable à d'autres méthodes classiques, quoique n'utilisant qu'un faible nombre de données en comparaison. / Generally, QSPR models are limited to individual compounds. In this thesis we have developed a QSPR approach to predict non-additive properties of binary mixtures, more explicitly their azeotropic behavior. To achieve this, several types of quantitative and qualitative models have been developed. This approach is original for two reasons. First, little research has been published on mixtures whose properties are no additive. Second, several new methodological aspects have been introduced in this work. First of all "special" descriptors able to describe mixtures have been proposed. In addition, a robust protocol for obtaining and validating models was used, and a reliable models applicability domain was proposed. The methodology developed during this thesis demonstrates the consistency of a new concept - the QSPR models for mixtures. It is comparable to other conventional methods, though using only limited data.

Submitted by Celso Magalhaes (celsomagalhaes@ufrrj.br) on 2017-06-14T14:33:49Z No. of bitstreams: 1 2016 - Camila de Souza Silva.pdf: 1802766 bytes, checksum: 154ae702e6dbdd0afdc13fb202a08682 (MD5) / Made available in DSpace on 2017-06-14T14:33:49Z (GMT). No. of bitstreams: 1 2016 - Camila de Souza Silva.pdf: 1802766 bytes, checksum: 154ae702e6dbdd0afdc13fb202a08682 (MD5) Previous issue date: 2016-08-02 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES / Equilibrium data at low and high pressures are important to correct establish conditions of temperature and pressure for separation processes, and to supply the capacity of the solvent, the compositions of the phases and the selectivity of the solvent. The separation of ethanol-water system is of great importance for the industry due to numerous applications of anhydrous ethanol. In all of these applications, ethanol must be free of water and it is necessary to add a third component in the distillation to break the azeotrope. It can be add different solvents, as benzene, hexane, ethyleneglycol, salts, and, in the last years, many studies have been done with ionic liquids. So, the purpose of this work is to use an ionic liquid (1-ethyl-3-methylimidazolium chloride) as the third component, looking for the ethanol dehydration. Because of that, it was done a study to evaluate the effect of this ionic liquid in the liquid-vapor equilibrium behavior between water and ethanol. Experimental data were measured, in triplicate, under normal pressure, in an Othmer-type ebulliometer (300 mL of volume), with two condensers, and made of borosilicate glass. The sample analysis was done in a digital densimeter. The ionic liquid used was recovered from one solution to another, just by adding the required amount to complete each mass fraction. Experimental data was measured with ethanol-water solutions varying the molar concentrations from 0.2 to 0.99, and ionic liquid weight fraction masses from 5 to 60%, to evaluate the behavior of the equilibrium data of the ethanol+water+[emim][Cl] system. The experiments showed that [emim][Cl] with a minimum mass fraction of 20% is a promising solvent because it could ?break? the azeotrope between water and ethanol, and higher mass fraction of ionic liquid were better to enrich the vapor phase in ethanol. NRTL model was used to correlate experimental vapor-liquid equilibrium of the ternary system, estimating the binary parameters, applying the bubble point methodology. The deviations of temperature and vapor phase composition were 0.147 ?C and 0.049, respectively. The relative volatility was greater than 1 for the mass fractions from 20%. The activity coefficients decrease with the increase in the molar concentration of ethanol. Values of the excess Gibbs free energy show a positive deviation for all mass fractions worked, and the experimental data were consistent thermodynamically / Os dados de equil?brio a press?es baixas e elevadas s?o importantes para estabelecer as condi??es corretas de press?o e temperatura para os processos de separa??o e para fornecer a capacidade do solvente, as composi??es das fases e a seletividade do solvente. A separa??o do sistema etanol-?gua ? de grande import?ncia para a ind?stria devido a numerosas aplica??es do etanol anidro. Em todas essas aplica??es, o etanol deve ser livre de ?gua e, para isso, ? necess?rio adicionar um terceiro componente na destila??o para quebrar o aze?tropo. Podem ser adicionados diferentes solventes como o benzeno, hexano, etilenoglicol, sais e, nos ?ltimos anos, tem-se visto muitos estudos com l?quidos i?nicos. Com isso, o objetivo deste trabalho ? a utiliza??o de um l?quido i?nico (1-etil-3-metil imidaz?lio cloreto) como terceiro componente, visando a desidrata??o do etanol, al?m da avalia??o do efeito deste l?quido i?nico no comportamento do equil?brio l?quido-vapor entre a ?gua e o etanol. Os dados experimentais foram medidos, em triplicata, sob press?o normal, em um ebuli?metro tipo Othmer (300 mL de volume), com dois condensadores, feitos de vidro de borosilicato. As determina??es das amostras foram feitas em um dens?metro digital. Os dados foram medidos com solu??es de etanol-?gua em diferentes concentra??es molares (0,2 a 0,95), variando a fra??o m?ssica de l?quido i?nico de 0,05 a 0,60, para avaliar o comportamento dos dados de equil?brio do sistema etanol-?gua-[emim][Cl].Os resultados mostraram que o [emim][Cl] ? um solvente promissor, pois "quebra" o aze?tropo entre a ?gua e etanol a partir de 20% de l?quido i?nico, e a concentra??o de etanol na fase vapor foi maior com o aumento da fra??o m?ssica de LI.O modelo NRTL foi utilizado para correlacionar os dados experimentais de equil?brio, estimando-se os par?metros bin?rios, aplicando-se a metodologia do ponto de bolha. Os desvios em rela??o ? temperatura e a composi??o molar da fase vapor foram 0,147 ?C e 0,049, respectivamente. O l?quido i?nico, recuperado de uma solu??o para outra, passou por uma an?lise de RMN para avaliar se n?o houve altera??o na sua estrutura e, constatou-se que, ap?s ser recuperado, e novamente reutilizado, o solvente n?o perdeu as caracter?sticas originais. As volatilidades relativas foram superiores a 1 para as fra??es m?ssicas a partir de 20%, confirmando a quebra do aze?tropo. A energia livre de Gibbs em excesso apresentou valores que mostram um desvio positivo para todas as fra??es m?ssicas trabalhadas e os dados experimentais foram consistentes termodinamicamente

No description available.

Este trabalho descreve a aplicação de redes neurais \"feed-forward\" com três camadas em diferentes áreas da Engenharia Química. O objetivo principal do projeto é a modelagem, simulação e posterior otimização do processo de produção de isopreno empregando técnicas de redes neurais em substituição as equações de modelagem fenomenológica. A planta industrial testada é a unidade de produção de isopreno da BRASKEM (antiga COPENE). O sistema consiste essencialmente de um reator de dimerização e uma série de colunas de destilação. Uma vez que redes neurais são capazes de aprender eficientemente o processo a partir de informações extraídas diretamente de dados da planta, para este trabalho o modelo de rede neural gerado foi construído a partir de dados históricos operacionais coletados a cada 15 minutos durante o período de 1 ano. Em uma primeira etapa é realizada a análise dos dados operacionais de modo a detectar e eliminar erros grosseiros e sistemáticos. Em seguida, a modelagem e simulação do processo são realizadas. O modelo de redes neurais gerado é, então, empregado na otimização qualitativa/quantitativa do processo, construindo um \"grid\" de busca detalhado da região de interesse, através um mapeamento completo da função objetivo no espaço das variáveis de decisão. A segunda etapa diz respeito à predição de azeótropos, visando um melhor entendimento do comportamento do sistema da seção de extração de isopreno. Nas duas etapas, a grande vantagem em utilizar modelos de redes neurais, além de ajustar dados, é a capacidade que estes apresentam em representar eficientemente sistemas multivariáveis, complexos e não lineares, aprendendo o sistema, sem o conhecimento das leis físicas e químicas que o regem. Comparações entre a predição dos modelos propostos e os dados experimentais foram executadas e resultados muito bons foram conseguidos do ponto de vista industrial. ) Esta metodologia fornece informações interessantes e de maior compreensão para a análise dos engenheiros de processo do que os procedimentos convencionais correspondentes. Além disso, este trabalho mostra que a metodologia de redes neurais é promissora para varias aplicações indústrias, tais como análise de dados, modelagem, simulação e otimização de processos, bem como predição de propriedades termodinâmicas. / This work describes the application of a three-layer feed-forward neural network (NN) in different areas of chemical engineering. The main objective of this study is to model, simulate and optimize a real industrial plant, using NN by replacing phenomenological models. The industrial process studied is the isoprene production unit from BRASKEM. The chemical process consists basically of a dimerization reactor and a separation column train. Since NNs are able to extract information from plant data in an efficient manner, for this work, the neural network model was built directly from historical plant data, which were collected every 15 minutes during a period of one year. These data were carefully analyzed in order to identify and eliminate gross error data and non-steady state operation data. The modeling using NN was carried out by parts in order to get information on intermediate streams. Then, the global model was built, by interconnecting each individual model, and used to simulate and optimize the process. The optimization procedure carries on a detailed grid search of the region of interest, by a full mapping of the objective function on the space of decision variables. The second stage of this work deals with the azeotropic prediction using also the neural network approach. The objective of this step was to obtain a better understanding of the system behavior in the isoprene extraction section. Since all the cases studied are non-linear, complex andmultivariable systems, NN approach appears as a technique of interest due to its capability of learning the system without knowledge of the physical and chemical laws that govern it. Comparisons between the model\'s prediction and the experimental data were performed and reasonable results were achieved from an industrial point of view. ) Using neural network approach provides more comprehensive information for an engineer\'s analysis than the conventional procedure. This work shows that the use of NN methodology is promising for several industrial applications, such as data analysis, modeling, simulation and optimization process, as well as thermodynamics properties prediction. However, success in obtaining a reliable and robust NN depends strongly on the choice of the variables involved, as well as the quality of available data set and the domain used for training purposes.

The convective drying of thin layers of multicomponentliquid mixtures into an inert gas, and the influence ofdifferent process controlling mechanisms on drying selectivityis studied. Drying experiments under gas-phase-controlledconditions are performed by low intensity evaporation, fromfree liquid surfaces, of ternary mixtures without non-volatilesolutes. Liquid-side-controlled experiments are carried out bydrying a multicomponent polymeric solution containing twovolatile components, one non-volatile polymer and an optionalnonvolatile softening substance. Mathematical models to describe gas- andliquid-side-controlled drying based on interactive diffusion inboth liquid and gas phases as the main mechanisms for masstransfer are developed. For gas-phase-controlled drying, astability analysis of the ordinary differential equations thatdescribes the evaporation process is performed. Isothermal andnon-isothermal drying processes are considered in batch andcontinuous modes. The mathematical model to describe thecomposition profiles during batch drying of the polymeric film,considering liquid resistance, is solved numerically. Due tothe lack of experimental data, properties for this polymericsystem are estimated by using established methods. Ananalytical solution of the diffusion equation, by assuming anisothermal drying process and a constant matrix ofmulticomponent diffusion coefficients is developed. For thecontinuous case, liquid-side resistance is studied by modellingevaporation of a multicomponent falling liquid film into aninert gas including indirect heating. The results of the gas-phase-controlled model are in goodagreement with experimental results. For the polymeric film,the agreement is only qualitative since the model does notaccount for a membrane that develops on the film surface. Thestability analysis permits the prediction of trajectories andfinal state of a liquid mixture in a gas-phase-controlleddrying process. For isothermal evaporation of ternary mixturesinto pure gas, the solutions are trajectories in the phaseplane represented by a triangular diagram of compositions. Thepredicted ternary dynamic azeotropic points are unstable orsaddle. On the other hand, binary azeotropes are stable whenthe combination of the selectivities of the correspondingcomponents is negative. In addition, pure component singularpoints are stable when they are contained within theirrespective isolated negative selectivity zones. Undernon-isothermal conditions, maximum temperature valuescharacterise stable azeotropes. Incremental loading of the gaswith one or more of the components leads to a node-saddlebifurcation, where a saddle azeotrope and a stable azeotropecoalesce and disappear. For continuous drying, the singularpoints are infinite and represent dynamic equilibrium pointswhose stability is mainly dependent on the ratio of inletgas-to-liquid flow rates. As long as the process isgas-phasecontrolled, these results also apply to a porous solidcontaining a liquid mixture. In general, liquid-side control makes the drying processless selective but it is difficult to maintain this conditionduring the whole process. Under the influence of its owndynamics, a process starting as liquid-side-controlled tendstowards a gas-phase-controlled process. The presence ofnon-volatile components and indirect heating may delay thisdevelopment. Considering the evolution of the processcontrolling steps and its influence on selectivity, a modelaimed at describing the complete trajectory of a drying orevaporation process must include the coexistence of allrelevant mechanisms. Keywords:ternary mixture, falling film, diffusionequation, gas-phase control, liquid-phase control, selectivity,stability analysis, polymeric solution, evaporation, azeotrope,batch drying, continuous drying.