Melting point finds applications in chemical identification, purification and in the calculation of a number of other physicochemical properties such as vapor pressure and aqueous solubility. Despite the availability of enormous amounts of experimental data, no generally applicable methods have been developed to estimate the melting point of a compound from its chemical structure. A quick estimation of melting point can be a useful tool in the design of new chemical entities.In this dissertation, a simple means of estimating the melting points for a large variety of pharmaceutically and environmentally relevant organic compounds is developed. Melting points are predicted from the separate calculation of the enthalpy and entropy of melting directly from the chemical structure. The entropy of melting is calculated using a semi-empirical equation based on only two non-additive molecular parameters. This equation is validated and refined using a large collection of experimental entropy of melting values. The enthalpy of melting is calculated by additive group contributions.Melting points are estimated from the ratio of the enthalpy of melting and the entropy of melting. All of the methods and group contributions developed in this study are compatible with the UPPER (Unified Physical Property Estimating Relationships) scheme. The predicted melting points are compared to experimental melting points for over 2200 organic compounds collected from the literature. The average absolute error in melting point prediction is 30.1 °. This is a very reasonable estimate considering the size and diversity of the dataset used in this study.

This thesis explores mathematical optimization techniques to address the computeraided molecular and mixture design problems (CAMD/CAMxD). In particular, we leverage the power of mixed-integer linear programs (MILPs) to quickly and efficiently design over the massive chemical search space. These MILPs, when coupled with state-ofthe- art derivative-free optimization (DFO) methods, make for an efficient optimization strategy when designing mixtures of molecules or when considering a single molecule design problem that involves difficult thermodynamics or process models. In the first chapter, we provide a very general overview of the field of CAMD as addressed from the perspective of mathematical optimization. We discuss many relevant quantitative structure-property relationships (QSPRs) and provide constraints typically used in CAMD/CAMxD optimization problems. The second chapter introduces our DFO-based molecular/mixture design algorithm and describes how this approach enables a much greater molecular diversity to be considered in the search space as compared to traditional methods. Additionally, this chapter looks at a few case studies relevant to crystallization solvents and provides a detailed comparison of 27 different DFO algorithms for solving these problems. The third chapter introduces COSMO-RS/-SAC as alternatives to UNIFAC as the method used to capture mixture thermodynamics for a variety of CAMD/CAMxD problems. To fully incorporate COSMO-RS/-SAC into CAMD, we introduce group contribution (GC) methods for estimating a few necessary parameters for COSMO-based methods. We demonstrate the utility of COSMO-RS/-SAC in a few case studies for which UNIFAC-like methods are insufficient. In the fourth chapter, we investigate reaction solvent design using COSMO-based methods. COSMO-RS is particularly suitable for these problems as they allow for modeling of many relevant species in chemical reactions (transition states, charges, etc.) directly at the quantum level. This information can be immediately passed to the CAMD problem. We investigate a number of solvent design problems for a few difficult reactions. We summarize the work and provide a few future directions in the final chapter. Overall, this thesis serves to push the field of CAMD forward by introducing new methods to more efficiently explore the massive chemical search space and to enable a few new classes of problems which were previously untenable.

Nesse trabalho, a equação de estado PC-SAFT é combinada com um método de contribuição de grupos (GC) para estimação dos seus parâmetros. Para tanto, foram utilizados dados experimentais dos componentes puros (pressão de vapor e volume específico do líquido) e em mistura (equilíbrio líquido-vapor - VLE e coeficiente de atividade em diluição infinita - IDAC). Através de uma análise de sensibilidade, verificou-se que o parâmetro volume de associação poderia ser mantido constante, reduzindo o número de parâmetros a serem estimados. O objetivo principal foi estudar misturas que apresentassem associação cruzada ou forte interação entre os compostos. Com os parâmetros estimados, avaliou-se o desempenho do modelo GC-PC-SAFT no cálculo de propriedades de n-alcanos, 1-álcoois, aminas, clorofórmio e acetona. Os desvios médios obtidos no cálculo do equilíbrio líquido-vapor (VLE), entre as diferentes misturas estudadas, mostraram que a estratégia adotada para a estimação do parâmetro energia de associação apresentou bons resultados, com desvios relativamente baixos para a maioria dos casos estudados. Para IDAC, as predições foram muito semelhantes àquelas obtidas por outros modelos. Os resultados de VLE são importantes, pois fornecem informações sobre as concentrações intermediárias de uma mistura, enquanto que o IDAC fornece uma medida eficiente do grau de não-idealidade da mistura. Essas propriedades foram escolhidas com o objetivo de conseguir uma melhor representação das misturas, buscando eliminar a necessidade de parâmetros de interação binária. Os resultados obtidos revelam que o modelo GC-PC-SAFT proposto pode ser utilizado para predizer o equilíbrio líquido-vapor com uma precisão satisfatória para sistemas binários entre os diferentes compostos estudados, sem nenhum parâmetro de interação binária. / In this work, the PC-SAFT EoS is combined with a group contribution method (GC) for parameter estimation. To achieve this, experimental data for pure components (vapor pressure and liquid volume) and mixtures (vapor-liquid equilibria - VLE and infinite dilution activity coefficient -IDAC) has been used. Through sensitivity analysis, it has been found that the association volume parameter could be set constant, thus reducing the amount of parameters that needed to be estimated. The aim of this work was to study mixtures that presented cross association or strong component interaction. With the estimated parameters, GC-PC-SAFT performance in properties calculation of n-alkanes, 1-alcohols, amines, chloroform and ketone has been evaluated. The average deviations obtained in the calculation of vapor-liquid equilibria (VLE), in the different mixtures considered, have shown that the strategy for association energy parameter estimation has presented good results, with relatively low deviations for most of the cases. For IDAC, the predictions presented very similar results to those obtained by other models. VLE results are important because they provide information about mixtures’ intermediary concentrations, whereas IDAC offers an efficient measure of mixtures’ degree of non-ideality. These properties have been chosen with the aim of getting a better representation of the mixtures, seeking to eliminate the need for binary interaction parameters. The obtained results show that GC-PC-SAFT can be used to predict vapor-liquid equilibria for binary systems among the different studied components with satisfactory accuracy with no binary interaction parameter.

<p>Background: In newspaper articles, we can nowadays read headlines as"Stop the robbing!"and"They pay extra tax through rent". A column in Göteborgsposten describes enormous amounts of money, pouring between municipal companies. Numerous of local newspapers have started to pay attention to their municipal companies and the transfer of profits from municipal housing enterprises and electric power companies. What are upsetting these journalists then? Both Hyresgästföreningen (The Swedish tenants’ association) and Boverket (The Swedish National Housing Board) indicate that the fiscal purpose of transactions within the group no longer is primary for municipalities. They allege that the transactions are improperly exploited by transferring inappropriate amounts of money from municipal housing enterprises. </p><p>Purpose: The purpose of this thesis is to map transactions between municipal group companies and also evaluate if they are consistent with the accounting regulation in the domain. Delimitation: The case study will have a geographical delimitation by only including the group of Linköpings Stadshus. The study will only include the years during 1997-2003. The transactions that will be examined are group contributions, shareholders’ contribution and dividend. The case study will also process the interest expense of the parent company because of the promissory note loan that was raised from the municipality of Linköping in connection with the formation of the group. </p><p>Realisation: In the form of a case study, the thesis was realised through the collection of annual reports, obtained from the companies concerned. These data was complemented through qualitative interviews with representatives from the parent company and the municipality. The complementation was made to increase the comprehension of the problems. </p><p>Results: The municipality is primarily using competition and focus in result as important factors when trying to imitate private companies. This does not harmonize with the purpose of the municipal transactions. It is also clear that municipal groups may be composing strategies dealing with empty the subsidiary companies out on capital. This applies particularly to electric power companies. In the case study, compared to a payment of tax, Tekniska Verken has not recovered enough shareholders’ contribution. With that, these companies financial position is clearly deteriorating compared to if they would pay tax. Regulations of housing enterprises were not created until 1999. That means that violations of the regulations were not possible before that either. Obviously, the groups were using this opportunity, and the dividend that has been discussed in the case study, is a clear example of the municipality not giving priority to the clients of the housing enterprise. Though, it is important to point out the legality of a dividend of that kind.</p>

Background: In newspaper articles, we can nowadays read headlines as"Stop the robbing!"and"They pay extra tax through rent". A column in Göteborgsposten describes enormous amounts of money, pouring between municipal companies. Numerous of local newspapers have started to pay attention to their municipal companies and the transfer of profits from municipal housing enterprises and electric power companies. What are upsetting these journalists then? Both Hyresgästföreningen (The Swedish tenants’ association) and Boverket (The Swedish National Housing Board) indicate that the fiscal purpose of transactions within the group no longer is primary for municipalities. They allege that the transactions are improperly exploited by transferring inappropriate amounts of money from municipal housing enterprises. Purpose: The purpose of this thesis is to map transactions between municipal group companies and also evaluate if they are consistent with the accounting regulation in the domain. Delimitation: The case study will have a geographical delimitation by only including the group of Linköpings Stadshus. The study will only include the years during 1997-2003. The transactions that will be examined are group contributions, shareholders’ contribution and dividend. The case study will also process the interest expense of the parent company because of the promissory note loan that was raised from the municipality of Linköping in connection with the formation of the group. Realisation: In the form of a case study, the thesis was realised through the collection of annual reports, obtained from the companies concerned. These data was complemented through qualitative interviews with representatives from the parent company and the municipality. The complementation was made to increase the comprehension of the problems. Results: The municipality is primarily using competition and focus in result as important factors when trying to imitate private companies. This does not harmonize with the purpose of the municipal transactions. It is also clear that municipal groups may be composing strategies dealing with empty the subsidiary companies out on capital. This applies particularly to electric power companies. In the case study, compared to a payment of tax, Tekniska Verken has not recovered enough shareholders’ contribution. With that, these companies financial position is clearly deteriorating compared to if they would pay tax. Regulations of housing enterprises were not created until 1999. That means that violations of the regulations were not possible before that either. Obviously, the groups were using this opportunity, and the dividend that has been discussed in the case study, is a clear example of the municipality not giving priority to the clients of the housing enterprise. Though, it is important to point out the legality of a dividend of that kind.

This work investigates the volatilities of amine solvents used in post-combustion CO₂ capture from coal-fired power plants. Amine volatility is one of the key criteria used in screening an amine solvent for CO₂ capture: (1) amine losses up the stack can react in the atmosphere to form ozone and other toxic compounds; (2) volatility losses can result in greater solvent make-up costs; (3) high losses will require the use of bigger water wash units, and more water, to capture fugitive amines prior to venting - these translate to higher capital and operating costs; (4) volatilities need to be measured and modeled in order to develop more accurate and robust thermodynamic models. In this work, volatility is measured using a hot gas FTIR which can determine amine, water, and CO₂ in the vapor headspace above a solution. The liquid solution is speciated by NMR (Nuclear Magnetic Resonance). There are two key contributions made by this research work: (1) it serves as one of the largest sources of experimental data available for amine-water volatility; (2) it provides amine volatility for loaded systems (where CO₂ is present) which is a unique measurement not previously reported in the literature. This work studied the volatility of 20 alkanolamines in water at 0.5 - 1.1 molal (m) in water (< 1.5 mol% amine) at zero loading (no CO₂) from 40 ° - 70 °C. An empirical group contribution model was developed to correlate H[subscript 'amine'] to molecular structures of both alkylamines and alkanolamines. The model incorporated additional functional groups to account for cyclic structures and to distinguish between different types of alkyl groups based on the attached neighboring groups. This model represented the experimental H[subscript 'amine'], which spanned five orders in magnitude, to well within an order of magnitude of the measured values. The second component of this research involves upgrading the AspenPlus® v.7.3 model of MDEA-PZ-CO₂-H₂O system primarily by improving MDEA thermodynamics for MDEA-H₂O, MDEA-CO₂-H₂O, and MDEA-PZ-CO₂-H₂O. A key modification was made to include the carbonate (CO₃²⁻) species into the model chemistry set which greatly improved the fit of CO₂ solubility for MDEA-CO₂-H₂O at ultra lean loading ([alpha]) for 0.001 < [alpha] < 0.01. With MDEA-PZ-H₂O, no MDEA-PZ cross interaction parameters were needed to match the blend volatility. Ultimately, both the blend volatility, at unloaded and loaded conditions, along with speciation were adequately represented by the upgraded model. The final component of this research involves screening the volatilities of novel amines at unloaded and nominal lean loading condition from 40 ° - 70 °C (absorber operating conditions). The volatility of tertiary and hindered amines, such as MDEA and AMP, respectively, is not a strong function of loading because these amines are unable to form stable carbamates. Conversely, the volatility of mono-amines and of diamines decreases by ~3 and 5-20 times, respectively, due to a much greater extent of carbamate-forming speciation. PZ or a blend having a diamine promoted by PZ would be favorable for CO₂ capture due to the low volatility of the diamines in loaded solution. . Finally, in order of increasing degree of salting out as reflected by the increasing magnitude of the system asymmetric amine activity coefficient, 7 m MDEA < 4.8 m AMP ~ 7 m MDEA/2 m PZ < 8 m PZ < 7 m MEA. / text

Nesse trabalho, a equação de estado PC-SAFT é combinada com um método de contribuição de grupos (GC) para estimação dos seus parâmetros. Para tanto, foram utilizados dados experimentais dos componentes puros (pressão de vapor e volume específico do líquido) e em mistura (equilíbrio líquido-vapor - VLE e coeficiente de atividade em diluição infinita - IDAC). Através de uma análise de sensibilidade, verificou-se que o parâmetro volume de associação poderia ser mantido constante, reduzindo o número de parâmetros a serem estimados. O objetivo principal foi estudar misturas que apresentassem associação cruzada ou forte interação entre os compostos. Com os parâmetros estimados, avaliou-se o desempenho do modelo GC-PC-SAFT no cálculo de propriedades de n-alcanos, 1-álcoois, aminas, clorofórmio e acetona. Os desvios médios obtidos no cálculo do equilíbrio líquido-vapor (VLE), entre as diferentes misturas estudadas, mostraram que a estratégia adotada para a estimação do parâmetro energia de associação apresentou bons resultados, com desvios relativamente baixos para a maioria dos casos estudados. Para IDAC, as predições foram muito semelhantes àquelas obtidas por outros modelos. Os resultados de VLE são importantes, pois fornecem informações sobre as concentrações intermediárias de uma mistura, enquanto que o IDAC fornece uma medida eficiente do grau de não-idealidade da mistura. Essas propriedades foram escolhidas com o objetivo de conseguir uma melhor representação das misturas, buscando eliminar a necessidade de parâmetros de interação binária. Os resultados obtidos revelam que o modelo GC-PC-SAFT proposto pode ser utilizado para predizer o equilíbrio líquido-vapor com uma precisão satisfatória para sistemas binários entre os diferentes compostos estudados, sem nenhum parâmetro de interação binária. / In this work, the PC-SAFT EoS is combined with a group contribution method (GC) for parameter estimation. To achieve this, experimental data for pure components (vapor pressure and liquid volume) and mixtures (vapor-liquid equilibria - VLE and infinite dilution activity coefficient -IDAC) has been used. Through sensitivity analysis, it has been found that the association volume parameter could be set constant, thus reducing the amount of parameters that needed to be estimated. The aim of this work was to study mixtures that presented cross association or strong component interaction. With the estimated parameters, GC-PC-SAFT performance in properties calculation of n-alkanes, 1-alcohols, amines, chloroform and ketone has been evaluated. The average deviations obtained in the calculation of vapor-liquid equilibria (VLE), in the different mixtures considered, have shown that the strategy for association energy parameter estimation has presented good results, with relatively low deviations for most of the cases. For IDAC, the predictions presented very similar results to those obtained by other models. VLE results are important because they provide information about mixtures’ intermediary concentrations, whereas IDAC offers an efficient measure of mixtures’ degree of non-ideality. These properties have been chosen with the aim of getting a better representation of the mixtures, seeking to eliminate the need for binary interaction parameters. The obtained results show that GC-PC-SAFT can be used to predict vapor-liquid equilibria for binary systems among the different studied components with satisfactory accuracy with no binary interaction parameter.

BUARQUE, Hugo Leonardo de Brito. Predição de propriedades de gasolinas a partir das suas composições. 2006. 206f. Tese (Doutorado em Física) - Curso de Pós-Graduação em Física, Universidade Federal do Ceará, Fortaleza, 2006. / Submitted by francisco lima (admir@ufc.br) on 2013-04-12T13:50:23Z No. of bitstreams: 1 2006_tese_hldbbuarque.pdf: 89844 bytes, checksum: 09dd0a3616e88ec1e47ab52519f63da5 (MD5) / Approved for entry into archive by Edvander Pires(edvanderpires@gmail.com) on 2014-02-25T20:59:40Z (GMT) No. of bitstreams: 1 2006_tese_hldbbuarque.pdf: 89844 bytes, checksum: 09dd0a3616e88ec1e47ab52519f63da5 (MD5) / Made available in DSpace on 2014-02-25T20:59:40Z (GMT). No. of bitstreams: 1 2006_tese_hldbbuarque.pdf: 89844 bytes, checksum: 09dd0a3616e88ec1e47ab52519f63da5 (MD5) Previous issue date: 2006 / Commercial gasolines are normally produced by blendin g hydrocarbon fractions obtained from the distillation of crude oil or from o ther petrochemical or refining processes, and carried through in order to comply with a variety of legal and ambient specifications at minimum cost. The quality for the use a nd commercialization of gasolines is evaluated through certain characteristics specified by governmental regulation. Such characteristics are usually determined by different methodologies and experimental techniques, since those depend on the ir constituents and their respective concentrations with a high complexity. Thus, blending of gasolines in petrochemical and refining industries is sometimes a very laborious procedure. The prediction of fuel properties from composition data is growing in importance in the last few years. Methods of group contribution have been usedin the last decades to predict properties of pure organic compounds and some mix ture parameters (e.g.,UNIFAC). However, most of the recent studies use artificial neural networks as a technique for prediction for fuel properties using the composition of classes of constituents or key-compounds as input data. The main a dvantage of a neural network is its capacity to extract general and unknown in formation for certain series of data (training), supplying useful and fast models for prediction. However, the use of neural networks trained to predict properties of fue ls produced from one given combination of petroleum fractions can not be suitable in the prediction of the characteristics of other gasolines produced from other orig ins due to the complexity and variability of gasoline composition. In this study, methods of multiple linear regression and artificial neural networks have been eval uated in the correlation and prediction of gasoline properties from information of composition obtained by gas chromatography, as well as a methodology for prediction of properties using a hybrid method composed of neural networks and group contribut ion. The developed model is evaluated and compared to other methods, revealing to be sufficiently promising for prediction of properties of pure components and com plex mixtures. / As gasolinas comerciais são normalmente produzidas a partir de combinações de frações oriundas da destilação do petróleo ou de outros processos petroquímicos e de refino e realizadas de modo a atender uma variedade de especificações legais e ambientais, com o mínimo de custo possível. A qualidade para o uso e comercialização de uma gasolina é avaliada através de cer tas características especificadas por leis e normas governamentais. Estas caracter ísticas são normalmente determinadas por diferentes metodologias e técnicas experimentais, haja vista que dependem dos seus constituintes e suas respecti vas concentrações com uma complexidade bastante elevada, tornando a formulação da gasolina originada em refinarias e petroquímicas, um procedime nto muitas vezes bastante laborioso. O intuito de se predizer propriedades de derivados de petróleo a partir de dados de composição é antigo e vem crescendo em importância nos últimos anos. Métodos de contribuição de grupos têm sido utilizados ao longo das últimas décadas para predizer propriedades de compostos orgânicos puros e alguns parâmetros de misturas (e.g., UNIFAC). Entretanto, a maior parte dos estudos mais recentes utiliza redes neurais artificiais como técnica para predição de propriedades de combustíveis usando a composição de grupos de compostos ou mesmo de compo stos-chave como informação de entrada. A principal vantagem de um a rede neural é sua capacidade de extrair informações gerais e desconhecidas pa ra certa série de dados (treinamento), fornecendo modelos de predição úteis e rápidos tanto para sistemas lineares como não-lineares. Porém, dada a complexidade e variabilidade dos constituintes das gasolinas, a utilização de redes neurais t reinadas para modelar as propriedades destes combustíveis produzidos a partir de uma dada combinação de frações petrolíferas pode não se adequar na predição da s características de gasolinas obtidas a partir de uma outra origem. Neste estudo, métodos de regressão linear múltipla e redes neurais artificiais foram avali ados na correlação e predição de propriedades de gasolinas a partir de informações de com posição obtidas por cromatografia gasosa, como também foi desenvolvida uma metodologia de predição de propriedades utilizando um método híbrido de redes neurais e contribuição de grupos. O modelo desenvolvido é avaliado e comparado aos demais, mostrando-se bastante promissor para predição de propriedades de componentes puros e misturas mais complexas.

AgÃncia Nacional do PetrÃleo / Commercial gasolines are normally produced by blending hydrocarbon fractions obtained from the distillation of crude oil or from other petrochemical or refining processes, and carried through in order to comply with a variety of legal and ambient specifications at minimum cost. The quality for the use and commercialization of gasolines is evaluated through certain characteristics specified by governmental regulation. Such characteristics are usually determined by different methodologies and experimental techniques, since those depend on their constituents and their respective concentrations with a high complexity. Thus, blending of gasolines in petrochemical and refining industries is sometimes a very laborious procedure. The prediction of fuel properties from composition data is growing in importance in the last few years. Methods of group contribution have been used in the last decades to predict properties of pure organic compounds and some mixture parameters (e.g., UNIFAC). However, most of the recent studies use artificial neural networks as a technique for prediction for fuel properties using the composition of classes of constituents or key-compounds as input data. The main advantage of a neural network is its capacity to extract general and unknown information for certain series of data (training), supplying useful and fast models for prediction. However, the use of neural networks trained to predict properties of fuels produced from one given combination of petroleum fractions can not be suitable in the prediction of the characteristics of other gasolines produced from other origins due to the complexity and variability of gasoline composition. In this study, methods of multiple linear regression and artificial neural networks have been evaluated in the correlation and prediction of gasoline properties from information of composition obtained by gas chromatography, as well as a methodology for prediction of properties using a hybrid method composed of neural networks and group contribution. The developed model is evaluated and compared to other methods, revealing to be sufficiently promising for prediction of properties of pure components and complex mixtures. / As gasolinas comerciais sÃo normalmente produzidas a partir de combinaÃÃes de fraÃÃes oriundas da destilaÃÃo do petrÃleo ou de outros processos petroquÃmicos e de refino e realizadas de modo a atender uma variedade de especificaÃÃes legais e ambientais, com o mÃnimo de custo possÃvel. A qualidade para o uso e comercializaÃÃo de uma gasolina à avaliada atravÃs de certas caracterÃsticas especificadas por leis e normas governamentais. Estas caracterÃsticas sÃo normalmente determinadas por diferentes metodologias e tÃcnicas experimentais, haja vista que dependem dos seus constituintes e suas respectivas concentraÃÃes com uma complexidade bastante elevada, tornando a formulaÃÃo da gasolina originada em refinarias e petroquÃmicas, um procedimento muitas vezes bastante laborioso. O intuito de se predizer propriedades de derivados de petrÃleo a partir de dados de composiÃÃo à antigo e vem crescendo em importÃncia nos Ãltimos anos. MÃtodos de contribuiÃÃo de grupos tÃm sido utilizados ao longo das Ãltimas dÃcadas para predizer propriedades de compostos orgÃnicos puros e alguns parÃmetros de misturas (e.g., UNIFAC). Entretanto, a maior parte dos estudos mais recentes utiliza redes neurais artificiais como tÃcnica para prediÃÃo de propriedades de combustÃveis usando a composiÃÃo de grupos de compostos ou mesmo de compostos-chave como informaÃÃo de entrada. A principal vantagem de uma rede neural à sua capacidade de extrair informaÃÃes gerais e desconhecidas para certa sÃrie de dados (treinamento), fornecendo modelos de prediÃÃo Ãteis e rÃpidos tanto para sistemas lineares como nÃo-lineares. PorÃm, dada a complexidade e variabilidade dos constituintes das gasolinas, a utilizaÃÃo de redes neurais treinadas para modelar as propriedades destes combustÃveis produzidos a partir de uma dada combinaÃÃo de fraÃÃes petrolÃferas pode nÃo se adequar na prediÃÃo das caracterÃsticas de gasolinas obtidas a partir de uma outra origem. Neste estudo, mÃtodos de regressÃo linear mÃltipla e redes neurais artificiais foram avaliados na correlaÃÃo e prediÃÃo de propriedades de gasolinas a partir de informaÃÃes de composiÃÃo obtidas por cromatografia gasosa, como tambÃm foi desenvolvida uma metodologia de prediÃÃo de propriedades utilizando um mÃtodo hÃbrido de redes neurais e contribuiÃÃo de grupos. O modelo desenvolvido à avaliado e comparado aos demais, mostrando-se bastante promissor para prediÃÃo de propriedades de componentes puros e misturas mais complexas.

No description available.