Domsjö Fabriker has earlier installed a Near Infrared (NIR) spectrometer after one of their recovery boilers. The purpose is to monitor the reduction efficiency of the boiler and later do process optimization. In this work calibration models for the instrument have been created. 108 green liquor samples have been extracted. A NIR spectrum was recorded for each sample and the samples were subsequently analyzed in laboratory for total alkali, sulfide and total sulfur. Several calibration models were created with multivariate data analysis and their performance and robustness were compared. The best model was able to predict reduction efficiency with a RMSEP of 2.75 percent units. Moreover, models were created for prediction of total alkali with a RMSEP of 0.108 mol/l, sulfides with a RMSEP of 1.95 g/l, total sulfur with a RMSEP of 2.83 g/l and S/Na2 ratio with a RMSEP of 0.022. The result is good enough that the instrument could be used to optimize the process and monitor process disturbances.

Submitted by Jean Medeiros (jeanletras@uepb.edu.br) on 2018-06-19T12:41:33Z No. of bitstreams: 1 PDF - Sheila Catarina de Oliveira.pdf: 372378827 bytes, checksum: 86635fa0b3f8f6691b4eb838b7559557 (MD5) / Approved for entry into archive by Secta BC (secta.csu.bc@uepb.edu.br) on 2018-07-03T17:46:30Z (GMT) No. of bitstreams: 1 PDF - Sheila Catarina de Oliveira.pdf: 372378827 bytes, checksum: 86635fa0b3f8f6691b4eb838b7559557 (MD5) / Made available in DSpace on 2018-07-03T17:46:42Z (GMT). No. of bitstreams: 1 PDF - Sheila Catarina de Oliveira.pdf: 372378827 bytes, checksum: 86635fa0b3f8f6691b4eb838b7559557 (MD5) Previous issue date: 2018-02-21 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Cachaça is a distilled beverage derived from the sugarcane juice. Considered one of the most consumed alcoholic beverages in Brazil and around the World, this alcoholic beverage has great influence in the economy of several Brazilian regions and it has attracted the interest of researchers from the most diverse science areas. The increase in the consumption and in the production of cachaça made it mandatory to establish a regulation regarding the quality control procedures. However, the methods currently applied have several disadvantages, such as being expensive and time-consuming. In this context, this dissertation aimed to develop an alternative method of analysis, using chemometric methods in association with infrared spectroscopy. The samples of commercial cachaça from different regions of Brazil were analyzed according to the standards established by the current legislation in the country regarding the physical-chemical properties: acidity, density, alcoholic content, lead and copper. Infrared spectroscopy spectra were collected in the range between 12000 and 4000 cm^-1 (NIR) and 4000 and 400 cm^-1(MIR). The unsupervised pattern recognition method of principal component analysis (PCA) was applied for an exploratory analysis of the data in order to identify patterns and clustering among the samples produced in the recently extinct Microregion of Brejo Paraibano. Partial least squares regression (PLSR) was used in the development of calibration models for the determination of physical-chemical parameters. All regression models constructed to predict the parameters of interest showed higher determination coefficients and smaller prediction errors in the spectral region of NIR. The values of R^2 , RMSECV and RMSEP were 0,977, 0,63 and 1,06 for the alcoholic content model and 0,960, 0,0014 e 0,0019 for the density model. The models constructed to predict the total acidity of the samples presented unsatisfactory results (R2=0,588, RMSECV=14,9074) and therefore were not used in the external calibration stage. These results demonstrate the potential of infrared spectroscopy analyzes for accurate prediction of density and alcohol contents of cachaça samples. Comparative tests using the MIR spectra resulted in higher values of values and lower determination coefficients. Therefore, it is possible to conclude that NIRS is a better option for the development of multivariate models of cachaça samples. Regarding the application PCA analysis, three assumptions were evaluated: there is a formation of defined groupings based on the producing region (using NIR and MIR); there is a formation of clusters according to alcoholic content limits (NIR) and finally; the samples can be grouped based on their physical-chemical results. The first two attempts did not present the formation of any group. However, the third application returned interesting results, allowing the grouping of samples in relation to their copper and calcium composition. Density and alcoholic concentration were also responsible for the discrimination of a group of samples. / A cachaça é uma bebida fermentodestilada oriunda do caldo de cana de açúcar. Considerada uma das bebidas alcoólicas de maior consumo no Brasil e no mundo, possui grande influência na economia de diversas regiões brasileiras e tem atraído cada vez mais o interesse de pesquisadores das mais diversas áreas. O aumento no consumo e na produção da cachaça tornou imprescindível o estabelecimento de uma legislação para o controle de qualidade da bebida. No entanto, os métodos utilizados atualmente apresentam inúmeras desvantagens, como uso de reagentes nocivos, alta demanda de tempo e investimento. Nesse contexto, essa dissertação visou o desenvolvimento de metodologias de análise alternativas, utilizando métodos quimiométricos em associação com a espectroscopia no infravermelho para o controle de identidade e de alguns parâmetros de qualidade do produto. As amostras de cachaça comerciais provenientes de diversas regiões do Brasil foram analisadas segundo os padrões de identidade e qualidade estabelecidos pela legislação vigente no país com relação aos parâmetros acidez, densidade, teor alcoólico, chumbo e cobre. Os dados de espectroscopia infravermelho foram obtidos na faixa entre 12000 e 4000 cm^-1 para o infravermelho próximo (NIR), e de 4000 e 200 cm^-1 para o infravermelho médio (MIR). O método de reconhecimento de padrões não-supervisionado de análises de componentes principais (PCA) e o método de regressão pelos mínimos quadrados parciais (PLSR), foram utilizados no desenvolvimento de métodos para a determinação dos teores dos parâmetros físico-químicos e na análise exploratória dos dados com o objetivo de identificar padrões de agrupamento e separação de amostras produzidas na recentemente extinta Microrregião do Brejo Paraibano. Todos os modelos de regressão construídos para a predição dos parâmetros de interesse apresentaram coeficientes de determinação mais elevados e menores erros de predição na região espectral do NIR das amostras de cachaça. Os valores de R^2, RMSECV e RMSEP foram de 0,977, 0,63 e 1,06 para o modelo do teor alcoólico e de 0,960, 0,0014 e 0,0019 para o modelo da densidade. Os modelos construídos para a predição da acidez total das amostras, apresentaram valores insatisfatórios (R =0,588, RMSECV=14,9074) e por isso não foram utilizado na etapa de calibração externa. Estes resultados demonstram a potencialidade das análises por espectroscopia de infravermelho NIR para predição acurada dos teores de densidade e álcool de amostras de cachaça. Os testes comparativos utilizando os espectros MIR resultou em altos valores de erro e baixos coeficientes de determinação, sendo possível concluir que a NIRS é uma melhor opção para o desenvolvimento de modelos multivariados de cachaças. Com relação a aplicação da análise de componentes principais foram avaliadas 3 situações: formação de agrupamentos definidos 2 baseados na região produtora (utilizando NIR e MIR); formação de agrupamentos de acordo com a conformidade ou não em relação ao teor alcoólico (NIR) e por fim a avaliação do comportamento das amostras com relação aos dados das análises de referência. As duas primeiras tentativas não apresentaram a formação de nenhum grupo. No entanto, a terceira aplicação retornou resultados interessantes, possibilitando o agrupamento de amostras em relação ao teor dos metais cobre e cálcio em sua composição.

Orientador: Celso Ulysses Davanzo / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-05T16:24:43Z (GMT). No. of bitstreams: 1 Lucato_MelissaUmata_D.pdf: 2256384 bytes, checksum: fd4564569486d21c7ee858263e6a0d83 (MD5) Previous issue date: 2005 / Resumo: O presente trabalho teve como principal objetivo a elaboração de metodologias analíticas que fornecessem informações diretas referentes aos problemas analisados, utilizando simultaneamente métodos não destrutivos de obtenção de espectros (FTIR) e calibração multivariada (PLS) com seleção de variáveis (AIgoritmo Genético e AIgoritmo das Projeções Sucessivas). Para efetivar tal objetivo foram estudados dois diferentes sistemas analíticos: I-Análise dos combustíveis comerciais (diesel, gasolina e álcool etílico automotivo): Foram analisadas propriedades de amostras comerciais dos combustíveis como: densidade, tipos de hidrocarbonetos, curva de destilação, propriedades de octanagem, entre vários outros, a partir de espectros de FTIR-ATR. Os valores de referência foram obtidos a partir de normas ASTM, que foram correlacionados com resultados da calibração multivariada através do método PLS e MLR, com e sem seleção de variáveis. Os erros obtidos apresentaram valores inferiores aos da reprodutibilidade do método padrão para a maioria das propriedades. Além disso, os resultados entre os modelos propostos foram comparados estatisticamente pelo testeF. II -Estudo das propriedades de degradação do PET, Poli(tereftalato de etileno): Foram determinadas a viscosidade intrínseca [h] e a concentração de grupos carboxílicos terminais (CEG) através da técnica de refletância difusa e técnica fotoacústica no infravermelho com calibração multivariada (PLS). Os modelos foram comparados estatisticamente pelo teste-F. Os resultados apontaram um método alternativo às laboriosas análises convencionais que empregam reagentes tóxicos, gerando resíduos / Abstract: This research project was mainly intended to elaborate analytical methodologies that supplied referring direct information to the analyzed problems, using simultaneously non destructive methods of attainment of spectroscopy (FTIR) and multivaried calibration (PLS) with variable selection (Genetic Algorithm and Successive Projections Algorithm). To accomplish this objective, two different analytical systems had been studied: 1- Analysis of commercial fuels (diesel, gasoline and automotive etílic alcohol): - Properties of commercial samples of fuels had been analyzed as: density, types of hydrocarbons, curve of destillation, octane number properties, among several others, from spectra of FTIR. The foundation values had been gotten from ASTM norms, that had been correlated with results of the multivariate calibration through PLS and MLR methods, using or not variable selection. The obtained errors had presented inferior values to the ones for reprodutibilidade of the standard method for the majority of the properties. Moreover, the results among the considered models were statistically compared by the F test . II- Study of the Properties of PET's degradation, Poly(ethylene therphtalate) : - Intrinsic viscosity [h] and the concentration of terminal carboxylic groups had been determined (CEG) by the technique of diffuse reflectance and photoacoustic technique in the infrared region with multivaried calibration (PLS). The models were statistically compared by the F test. The results point to an alternative method for the laborious conventional analyses that use toxic reagents and generates residues / Doutorado / Quimica Analitica / Doutor em Ciências

Orientadores: Ronei Jesus Poppi, Cesar Alexandre de Mello / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-14T13:51:34Z (GMT). No. of bitstreams: 1 Ribeiro_DiorginisBuenoMontrazi_D.pdf: 1788617 bytes, checksum: f8b9685d4131466922bc03a5fa000661 (MD5) Previous issue date: 2009 / Resumo: Nesta tese metodologias analíticas foram desenvolvidas empregando espectroscopia Raman de baixa resolução amplificada por superfície (SERS) e calibração multivariada baseada no método dos mínimos quadrados parciais (PLS) para determinação dos pesticidas endosulfan e metamidofos (e misturas deles) em água, e do hormônio tireoestimulante (TSH) em plasma. Para a construção dos modelos de calibração dos pesticidas, um total de 70 e 30 amostras compuseram os conjuntos de calibração e validação, respectivamente, sendo a divisão realizada pelo algoritmo de Kennard-Stone. Já na construção dos modelos de calibração para a determinação quantitativa dos pesticidas nas misturas, um total de 38 e 11 amostras foram utilizadas na calibração e validação respectivamente. Para construção dos modelos de calibração para quantificação de TSH, um total de 39 mostras foram utilizadas na calibração e 14 amostras de plasma foram utilizadas na validação. As amostras também foram divididas pelo algoritmo de Kennard-Stone. Os modelos foram desenvolvidos utilizando diferentes tipos de pré-processamentos de sinais e comparados através dos erros de previsão (RMSEP). Foram utilizados como pré-processamento, o filtro de transformada de Fourier, a correção de espalhamento multiplicativa, a transformação padrão normal de variação, a ortogonalização de espectros pelo metodo de Gram-Schmidt, a centralização e autoescalamento dos dados. Os melhores modelos foram validados através da determinação de figuras de merito. Foram avaliados a exatidão, sensibilidade, sensibilidade analítica, seletividade, ajuste, razão sinal/ruído, limites de detecção e quantificação. A metodologia proposta mostrou-se rápida, de baixo custo e apresentou erros abaixo de 10mg/L para os pesticidas e abaixo de 0,8 mUI/mL para TSH, podendo facilmente ser adaptada para o monitoramento de pesticidas em águas e também em análises laboratoriais de rotina para determinação de TSH. / Abstract: In this thesis analytical methodologies were developed employing low resolution surface enhanced Raman spectroscopy (SERS) and multivariate calibration based on partial least squares method (PLS) for determination of the pesticides endosulfan and methamidophos (and mixtures of them) in water and the thyroid stimulating hormone (TSH) in plasma. For the pesticides calibration model development, a total of 70 and 30 samples composed the calibration and validation sets, respectively, using the Kennard-Stone algorithm for samples separation. In the model development for the mixture of pesticides, a total of 38 and 11 samples were used in the calibration and validation sets, respectively. For the model development in the TSH determination, 39 samples were used in the calibration set and 14 in the validation set. Also the Kennard-Stone algorithm was used to split the samples into the two data sets. The models were developed using different preprocessing methods and compared by using the prediction errors (RMSEP). The following pre-processing were tested: Fourier transform filter, multiplicative scatter correction, standard normal variate, spectra orthogonalization by Gram-Schmidt method, mean center and autoscaling. The best models were validated by figures of merit determination. It was assessed the accuracy, sensibility, analytical sensibility, selectivity, fit, signal/noise ratio, detection and quantification limits. The proposed methodology is fast, has low cost and presented prediction errors below to 10 mg/L for the pesticides and below to 0.8mìUI/mL for TSH. It may easily be adapted for the pesticides monitoring in waters and also for routine laboratory analysis in TSH determination. / Doutorado / Quimica Analitica / Doutor em Ciências

There is a need for a real-time, non-invasive method to monitor blood pH in a patient line during cardiopulmonary bypass, as today’s methods are both invasive and time consuming. Blood pH is an indicator of physiological and biochemical activity in the body and needs to be kept within a relatively narrow range, typically between 7.35-7.45. A pH value outside this range can be critical for the patient and therefore needs to be carefully monitored throughout the course of cardiopulmonary bypass. In this study the feasibility of using spectroscopic methods for indirect measurement of pH was investigated, and both transmission and reflectance spectroscopy were tested. The results showed that NIR reflectance spectroscopy is a feasible technique for blood pH monitoring during cardiopulmonary bypass. A strong correlation was found between measured pH values and spectral output in the wavelength range 800-930 nm. It was suggested that by means of the statistical partial least square regression method, a model could be created with three regression factors with a cross-validated R2 of 0.906 and a prediction error RMSEP of 0.089 pH units. The results presented here form a foundation for further analysis and experiments with larger sample set and more controlled experimental environment.

Flow injection systems have great potential for sample handling and analysis in process analytical chemistry. The flexibility and versatility of flow injection manifolds can he utilized in specific applications of sample conditioning and analysis. An overview of various flow injection methods, including flow reversals, double injection, and sequential injection is given, as well as different clean-up methods, such as gas diffusion, solid phase extraction, dialysis, and solvent extraction. Calibration techniques, such as single standard and multivariate calibration are also discussed. In addition, different aspects of process analytical chemistry, in particular sampling and sample handling, are discussed. The papers in this thesis describe a number of flow systems, where gradient-, gas diffusion-, and solid phase extraction- methodologies are applied, all of which have potential use in process analytical chemistry. Paper I is focused on multicomponent analysis of mixtures of organic acids by mathematically extracting information from complex spectra. The selectivity is improved by generating pH-gradients in the flow system. In paper II, the methodology of sensor injection is described and electrochemical and spectroscopic sensors are implemented in a sequential injection system. The method is illustrated by using pH sensors and a glucose electrode. Ammonia and ammonium ions are determined on-line to a bioprocess by gas diffusion in paper III. The benefit of frequent re-calibrations and in-line cleaning sequences are demonstrated. Finally a method for on-line determination of o-diphenols in the kraft process has been developed (paper IV). The o-diphenols are isolated from black liquor samples by solid phase extraction and thereafter transferred to a high performance liquid chromatography system for separation and quantification. / <p>Diss. (sammanfattning) Umeå : Umeå universitet, 1996, Härtill 4 uppsatser</p> / digitalisering@umu

La variété des climats et des sols du Brésil assure une production de fruits diversifiée au cours de toutes les saisons de l'année. Ainsi, le pays gagne en importance dans le domaine de l'horticulture. Le fruit de la passion est un fruit dont la production s’est développée au cours des dernières années dans et hors du pays en raison de l'augmentation de la consommation de fruits frais et de produits dérivés de la pulpe. Pour le meilleur usage des fruits, il est nécessaire de connaître les principaux paramètres de qualité. Les fruits sont des produits hautement périssables et dont le métabolisme se poursuit après la récolte. Il est donc important de pouvoir proposer des méthodes permettant un contrôle rapide et efficace, afin de les qualifier de manière plus précise, que ce soit pour la consommation, l'exportation ou la transformation. Les méthodes traditionnelles utilisées pour déterminer les paramètres de qualité des fruits sont essentielles pour la prévision des rendements potentiels dans les procédés industriels et également dans l'établissement de la valeur commerciale du fruit, cependant, elles sont coûteuses, souvent longues, demandent une main d’œuvre formée et sont en règle généralement destructrices. En ce sens, l’objectif de cette étude était de vérifier l'efficacité de l'application de spectroscopie dans les gammes du proche et du moyen infrarouge associée à une calibration multivariée pour déterminer les paramètres de qualité des fruits passion intacts, ainsi que des pulpes fraîches et lyophilisées. Les méthodes multivariées utilisées ont été l’analyse en composantes principales (ACP) et la régression linéaire par la méthode des moindres carrés partiels («partial least squares» ou PLS). Les paramètres de qualité analysés ont été les concentrations en sucres simples, les solides solubles totaux (°Brix), l’acidité totale, les concentrations des principaux acides organiques, la teneur en vitamine C et en caroténoïdes. Une faible corrélation a été obtenue entre les valeurs de référence et les valeurs calculées à partir des modèles PLS pour déterminer les concentrations des paramètres de la qualité du fruit intact. Par contre des prédictions satisfaisantes ont été obtenues à partir des spectres moyen-Infrarouge de la pulpe pour la détermination des concentrations en glucose, fructose, saccharose, AT et SSC. La région du proche infrarouge a donné des résultats moins bons à cause de la prédominance des bandes d'absorption de l'eau dans cette région. Par ailleurs, des modèles multivariés développés pour prédire les paramètres de qualité de la pulpe lyophilisée ont donné d'excellents résultats en particulier dans la région du moyen infrarouge pour le glucose (R2V = 0,938), le fructose (R2V = 0,907), le saccharose (R2V = 0,921) et l'acide citrique (R2V = 0,918). Les résultats pour l'acide malique et les caroténoïdes ne sont satisfaisants pour aucune des deux régions, probablement à cause des faibles concentrations de ces composés. Il a été constaté que: i) il est plus diffiicle d’obenir des équations de prédiction pour les composés en faible concentration; ii) le rayonnement infrarouge pénètre peu dans les fruits (quelques mm), et donc l’existence d’un épiderme épais constitue un obstacle à l'utilisation de cette technique, et iii) l'interférence de l'eau dans l'identification des bandes et des groupes spécifiques a été significative, à cause de la haute teneur en humidité des fruits. La région du moyen infrarouge semble mieux appropriée pour prédire la concentration des paramètres de qualité que le proche infrarouge. Les meilleurs résultats ont été obtenus en utilisant la gamme du moyen infrarouge et de la pulpe fraîche, de bonnes corrélations pouvant être obtenues sans l'étape de lyophilisation de la pulpe / Climate and soil variety of Brazil ensures a diversified fruit production during all seasons. Thus the country is gaining prominence in the orcharding field. Passion fruit is a fruit that stood out in recent years internally and out of the country due to the increased consumption of fresh fruit and products derived from the pulp. For the best use of this fruit, it is necessary to determine its main quality parameters. Since passion fruits are highly perishable products and their metabolism proceeds even after harvest, to establish methodologies that allow fast and efficient control becomes essential to qualify them in a more adequate way, either for export, processing or fresh consumption. Traditional methodologies employed to determine  fruit quality parameters have been fundamental to prediction of potential yields in industrial processes and also to establish commercial value of the fruit, however, they are expensive, time consuming and destructive. Therefore, the aim of this study was to assess the efficiency of mid and near infrared spectroscopy coupled to multivariate calibration to determine quality parameters of the intact passion fruit, as well as fresh and lyophilized pulps. Quality parameters analyzed were simple sugars, SSC, AT, organic acids, vitamin C and carotenoids. Results showed low correlations between reference analyses values and predicted values to determine the quality parameters concentrations in intact fruit. Multivariate models developed to quality parameters of in nature pulp showed good prediction results to determine the concentration of glucose, fructose, sucrose, AT, and SSC using mid-Infrared. NIR region showed higher prediction difficulty due to the higher water absorption bands in this region. on the other hand, multivariate models developed to predict quality parameters of the lyophilized pulp presented excellent results especially in the MID region  for glucose (R2v= 0,938), fructose (R2v = 0,907), sucrose (R2v = 0,921) and citric acid (R2v = 0,918). The results for malic acid and carotenoids contents were unsatisfactory in the two regions used in this study due to the low concentration of these compounds. From the results obtained in this study, it could be verified that: i) there is greater difficulty to correlate constituents at low concentrations, ii) light intensity through the fruit decreases significantly with skin depth, so thick skin has showed to be a barrier for using this technique and iii) water interference in identifying groups and specific bands was significant for passion fruit, due to its high moisture content. The results also revealed that the mid-Infrared region was most suitable for predicting the concentration of quality parameters, in comparison with NIR region. Based also on the results obtained, the use of fresh pulp with MID spectroscopy seemed to be the most suitable choice for quality parameters determination of the passion fruit, since good correlations could be obtained without pulp lyophilization step

Estudou-se o processo de absorção e dessorção de CO2 em solução aquosa da mistura de metildietanolamina (MDEA) e piperazina (PZ). Os ensaios de absorção foram realizados numa coluna de parede molhada com promotor de película, e, os ensaios de dessorção num sistema de semibatelada, ambos em escala de laboratório. Os testes experimentais de absorção foram realizados a 298 K e pressão atmosférica, com vazão de gás (CO2 e ar atmosférico) de 2,2.10-4 m3 s-1 e as seguintes vazões de líquido: 1,0.10-6; 1,3.10-6 e 1,7.10-6 m3 s-1. O sistema de absorção foi caracterizado através da determinação da área interfacial, a, o coeficiente volumétrico de transferência de massa, kGa, e o coeficiente volumétrico global médio de transferência de massa, KGa. No caso dos ensaios de dessorção, estes foram realizados nas temperaturas de 353, 363 e 368 K, onde empregou-se uma solução carbonatada de 10% PZ-20% MDEA e uma corrente de ar atmosférico nas vazões de 1,1.10-5 m3 s-1 e 2,7.10-5 m3 s-1. Este sistema foi caracterizado através da determinação do coeficiente volumétrico global de transferência de massa, KLa. Os resultados experimentais da área interfacial mostram que este é função da vazão do líquido, sugerindo uma maior área de irrigação como o aumento desta, onde teve-se uma maior área de transferência de massa. O resultado do parâmetro, KGa, indica uma dependência da vazão de líquido, a qual está associada à variação da área interfacial e à dependência do parâmetro KG com o perfil das concentrações da MDEA e PZ ao longo da coluna. A partir da teoria do duplo filme e pelo conhecimento dos parâmetros KGa, a e kGa, estimou-se um parâmetro cinético-difusivo associado à fase líquida, (( ) ) . Os resultados experimentais mostram que esse parâmetro varia pouco com a vazão de líquido, indicando tratar-se de um processo independente da hidrodinâmica do líquido, característico de sistemas com reação rápida. A concentração das aminas e carbamatos, nos ensaios de absorção e dessorção, foi determinada através dos modelos de calibração obtidas pela técnica de espectroscopia no infravermelho. Nos ensaios de absorção, foram observados que a concentração de PZ teve uma variação considerável (4 a 5% massa massa-1), entanto que a de MDEA variou pouco (0,3 a 0,5% massa massa-1), sugerindo que o processo de absorção de CO2 na mistura MDEA-PZ é controlado principalmente pela PZ, e supõe-se que a MDEA tem um papel de receptor de prótons procedentes da reação entre a PZ e o CO2. Nos ensaios de dessorção, observou-se que esse processo é afetado pela temperatura, sendo que, em temperaturas perto da ebulição (372 K), a taxa de dessorção de CO2 é maior do que em temperaturas menores, em certa forma é devido à dependência da velocidade de reação química com a temperatura. Os resultados do parâmetro KLa indicam que este diminui em função da concentração de carbamato de PZ (por exemplo, na temperatura de 368 K, de 7,5.10-4 a 1,0.10-4 s-1), devido a que este componente é decomposto em altas temperaturas gerando o CO2 e as aminas, sugerindo uma diminuição na velocidade de dessorção de CO2. Assim também, os resultados experimentais do parâmetro KLa indicam que este aumenta ligeiramente com a vazão do gás. / The CO2 absorption and desorption process was studied in an aqueous mixture of methyldiethanolamine (MDEA) and piperazine (PZ). The absorption process was carried out in a wetted wall column with film promoter sized with inner diameter of 0,0184 m and height of 0,67 m. The desorption tests were carried out in semi batch system at laboratory scale. The absorption experimental tests were conducted at 298 K, atmospheric pressure, gas flow rate of 2,2.10-4 m3 s-1 and the following liquid flow rates: 1,0.10-6; 1,3.10-6 and 1,7.10-6 m3 s-1. The absorption process was characterized by the following mass transfer parameters: interfacial area, a, the volumetric mass-transfer coefficient of the gas phase, kGa, and the overall volumetric mass-transfer coefficient, KGa. In the case of desorption tests, these were carried out at temperatures of 353, 363 and 368 K, where was used a carbonated solution of 10% PZ-20% MDEA (% wt) and the air flow rates of 1,1.10-5 m3 s-1 and 2,7.10-5 m3 s-1. The desorption process was characterized by the overall volumetric mass-transfer coefficient of the liquid phase, KLa. The experimental results of the interfacial area indicate that this parameter increases in function of the liquid flow rate, this effect is due to an larger irrigation on the inner wall of the column, and hence an available larger contact surface for the mass transfer. The overall volumetric mass-transfer coefficient, KGa, is influenced by the liquid flow rate, which is associated with a change in the interfacial area and the dependence of the parameter KG of the concentration profile of MDEA and PZ along the column. From the two-film theory and by knowledge of the parameters \"kGa\", \"a\" and \"KGa\", it was estimated the kinetic-diffusive parameter, (( ) ) . The experimental results show that this parameter varies slightly in function of the liquid flow rate, and this behavior is characteristic of systems with rapid reaction. Furthermore, this behavior suggests that the absorption process is independent of liquid hydrodynamic occurring in the column. The concentrations of MDEA, PZ and PZ carbamate, in the absorption and desorption tests, were determined by multivariate calibration models based on mid-infrared spectroscopy. In the absorption test, it was observed that the PZ concentration had a considerable variation (4 to 5% w w-1), whereas MDEA varied slightly along the column (0,3 to 0,5% w w-1), suggesting that the CO2 absorption process in the MDEA/PZ blend is mainly controlled by PZ, and it is supposed that MDEA participates as a proton receptor coming from reaction between CO2 and PZ. The results of the desorption tests shown that this process is affected by desorption temperature, and at temperatures near the boiling point (372 K), the CO2 desorption rate is higher than at lower temperatures, a certain shape is due to the dependence chemical reaction rate with temperature. The results of parameter KLa indicated that this decreases in function of the concentration of PZ carbamate (for instance, at the temperature of 368 K, from 7,5.10-4 to 1,0.10-4 s-1), because this component is decomposed at high temperatures generating CO2 and amines, suggesting a decrease in the CO2 desorption rate. Likewise, experimental results of the parameter KLa indicated that this increased slightly with the gas flow rate.

Submitted by Cássia Santos (cassia.bcufg@gmail.com) on 2014-09-19T11:19:07Z No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-09-19T11:25:02Z (GMT) No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-09-19T11:25:02Z (GMT). No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2013-05-03 / This work proposes the use of multi-objective genetics algorithms NSGA-II and SPEA-II on the variable selection in multivariate calibration problems. These algorithms are used for selecting variables for a Multiple Linear Regression (MLR) by two conflicting objectives: the prediction error and the used variables number in MLR. For the case study are used wheat data obtained by NIR spectrometry with the objective for determining a variable subgroup with information about protein concentration. The results of traditional techniques of multivariate calibration as the Partial Least Square (PLS) and Successive Projection Algorithm (SPA) for MLR are presents for comparisons. The obtained results showed that the proposed approach obtained better results when compared with a monoobjective evolutionary algorithm and with traditional techniques of multivariate calibration. / Este trabalho propõe a utilização dos algoritmos genéticos multiobjetivo NSGA-II e SPEA-II na seleção de variáveis em problemas de calibração multivariada. Esses algoritmos são utilizados para selecionar variáveis para Regressão Linear Múltipla (MLR) com dois objetivos conflitantes: o erro de predição e do número de variáveis utilizadas na MLR. Para o estudo de caso são usado dados de trigo obtidos por espectrometria NIR com o objetivo de determinar um subgrupo de variáveis com informações sobre a concentração de proteína. Os resultados das técnicas tradicionais de calibração multivariada como dos Mínimos Quadrados Parciais (PLS) e Algoritmo de Projeções Sucessivas (APS) para a MLR estão presentes para comparações. Os resultados obtidos mostraram que a abordagem proposta obteve melhores resultados quando comparado com um algoritmo evolutivo monoobjetivo e com as técnicas tradicionais de calibração multivariada.

<p>This thesis is a part of a project where a new method, the base descriptor approach, is studied. The purpose of this method is to reduce drift and extract vital information from electronic tongue measurements. Reference solutions, called descriptors, are measured and the measurements are used to find base descriptors. A base descriptor is, in this thesis, a regression vector for prediction of the property that the descriptor represent. The property is in this case the concentration of a chemical substance in the descriptor solution. Measurements from test samples, in this case fruit juices, are projected onto the base descriptors to extract vital and drift-stable information from the test samples. </p><p>The base descriptors are used to determine the concentrations of the descriptors'chemical substances in the juices and thereby also to classify the different juices. It is assumed that the measurements of samples of juices and descriptors drift the same way. This assumption has to be true in order for the base descriptor approach to work. The base descriptors are calculated by multivariate regression methods like partial least squares regression (PLSR) and principal component regression (PCR). </p><p>Only two of the descriptors tested in this thesis worked as basis for base descriptors. The base descriptors'predictions of the concentrations of chemical substances in the juices are hard to evaluate since the true concentrations are unknown. Comparing the projections of juice measurements onto the base descriptors with a classification model on the juice measurements performed by principal component analysis (PCA), there is no significant difference in drift of the juice measurements in the results of the two methods. The base descriptors, however, separates the juices for classification somewhat better than the classification of juices performed by PCA.</p>