The labelling of methyl tertiary butyl ether (MTBE), an oxygenate additive used extensively in gasoline blending, as an environmentally harmful chemical has led to the banning and subsequent phasing-out of this additive in California (USA). In response, the global petroleum industry is currently considering replacement strategies, which include the use of tertiary amyl methyl ether (TAME) or ethanol. Subsequently, SASOL (South African Coal and Oil Limited), a local petrochemical company, in its capacity as an environmentally responsible player in the global petroleum and aligned chemical markets, has commissioned this investigation into the environmental fate of the fuel oxygenates: TAME, ethanol and MTBE. In order to evaluate the environmental fate of the oxygenates, this dissertation has formed a three-tiered approach, using MTBE as a benchmark. The first tier assessed the general fate behaviour of the oxygenates using an evaluative model. A generic evaluative model, developed by Mackay et al. (l996a), called the Equilibrium Criterion (EQc) model was used for this purpose. This fugacity based multimedia model showed MTBE and TAME to have similar affinities for the water compartment. Ethanol was demonstrated to have a pre-disposition for the air compartment. Parameterisation of the EQC model to South African conditions resulted in the development of ChemSA, which reiterated the EQC findings. The second tier quantified the persistence (P), bioaccumulation (B) and long-range transport (LRT) potential of the additives. This tier also included a brief toxicity (T) review. MTBE and ethanol were demonstrated to be persistent and non-persistent, respectively, according to three threshold limit protocols (Convention on the Long Range Trans-boundary Air Pollution Persistent Organic Chemical Protocol; the United Nations Environment Programme Global Initiative; and the Track 1 criteria as defined by the Canadian Toxic Substances Management Policy, as referred to by the Canadian Environmental Protection Act 1999). These protocols were not unanimous in the persistence classification of TAME. Further investigation of persistence was conducted using a persistence and long-range transport multimedia model, called TaPL3, developed by Webster et al. (1998) and extended by Beyer et al. (2000). TaPL3 reiterated the conclusions drawn from the threshold limit protocols, indicating that TAME's classification worsened from non-persistent to persistent on moving from an air emission to a water emission scenario. This served to emphasise the negative water compartment affinity associated with TAME. Using classification intervals defined by Beyer et al. (2000), TaPL3 demonstrated that the long-range transport potential of the oxygenates increased in the order of TAME, ethanol and MTBE; however, it was concluded that none of the oxygenates were expected to pose a serious long-range transport threat. Bioaccumulation was not expected to be a pertinent environmental hazard. As expected, the oxygenates were dismissed as potential bioaccumulators by the first level of a screening method developed by Mackay and Fraser (2000); as well as by the threshold limit protocols listed above. Simulation of biomagnification, using an equilibrium food chain model developed by Thomann (1989), demonstrated that none of the oxygenates posed a biomagnification threat. A review of toxicity data confirmed that none of the three oxygenates are considered particularly toxic. LDso values indicated the following order of increasing toxicity: ethanol, MTBE and TAME. The third tier focussed on oxygenate aqueous behaviour. A simple equilibrium groundwater model was used to analyse the mobility of the oxygenates in groundwater. TAME was found to be 21 % less mobile than MTBE. Ethanol was shown to be very mobile; however, the applicability of the equilibrium model to this biodegradable alcohol was limited. An analysis of liquid-liquid equilibria comprised of oxygenate, water and a fuel substitution chemical was performed to investigate fuel-aqueous phase partitioning and the co-solvency effects of the oxygenates. Ethanol was shown to partition appreciably into an associated water phase from a fuel-phase. Moreover, this alcohol was shown to act as a co-solvent drawing fuel chemicals into the water phase. MTBE was found to partition sparingly into the water phase from a fuel-phase, with TAME partitioning less than MTBE. Neither ether was shown to act as a co-solvent. It was concluded that TAME and ethanol pose less of a burden to the environment than MTBE. Ethanol was assessed to be environmentally benign; however, it was concluded that ethanol's air compartment affinity and the extent of its co-influence on secondary solutes justified the need for further investigation before its adoption as a fuel additive. This project showed sufficient variation in the environmental behaviour of TAME and MTBE to justify the abandonment of the axiom that MTBE and TAME behave similarly in the environment. However, as MTBE is a significant water pollutant, and TAME has been shown to share a similar water affinity, it is cautiously recommended that the assumption of environmental similarity be discarded, except for the water compartment. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2002.

Gasoline Partially Premixed Combustion (PPC) was investigated in a single cylinder optical diesel engine. The PPC operation was achieved with a combination of high dilution and higher intake charge temperature at part-load conditions using Primary Reference Fuel (PRF). The relative air/fuel ratio (λ) was set to 2.3 and the EGR rate at 22%. Split injections of three fuel distribution strategies (50:50, 70:30 and 30:70) were studied. In addition, the effect of injection pressure (900 and 1200 bar) was investigated for each injection timing. The emission and performance of the gasoline PPC operations were then compared with those of the baseline diesel combustion operation. Based on the thermodynamic analysis of the engine performance, detailed in-cylinder studies were carried out by means of optical techniques. The high speed imaging technique was employed to observe the fuel spray development and combustion processes. A simultaneous Mie-LIF technique was then developed and utilized for the visualization of fuel liquid and vapour formation.

This thesis analyzed Georgia's pre and post embargo consumption of gasoline and residential electricity from 1960 to 1982 to determine: 1) if the structure of gasoline demand was stable; 2) to investigate the arguments for conservation; and 3) to ascertain that the pricing mechanism is indeed able to adjust consumption to levels of supply. Elasticities were computed and they were used to address simple tax issues. The study was significant for several reasons: 1) it is the first ever known study of its kind done on Georgia; 2) Georgia has a ninety five percent dependence rate on other states and foreign countries (indirectly) for her supply of energy resources; 3) the impact of the oil embargo on consumption, prices and the economy at large is necessary both as a post moterm and for future policy decisions. The results overwhelmingly favored stability in the structure of these demands. As such, the restricted model was valid for any projections and conclusions. The short-term price and income elasticities for gasoline were .30 and .06 and the long-term coefficients were 1.5 and .20 respectively. For electricity, the short-term price and income elasticities were .22 and .37 and the long-term estimates were .81 and 1.33 respectively. The weather variables were very inelastic. All coefficients were significant by the usual criteria. With inelastic demands for these energy resources, taxation seemed a welcome proposition for controlling prices and consumption and for generating revenue. However, it is only to the extent that other monetary, fiscal and economic objectives of the government are not jeopardized. Thus, the arguments for conservation holds to a good degree. A hybrid of a comprehensive tax-pricing policy and conservation efforts is necessary for stability in the energy sector.

As the legislation on vehicle emissions is becoming more and more stringent, increasing attention has been paid to the fine particles emitted by diesel and gasoline vehicles. The high number emission of fine particles has been shown to have a large impact on the atmospheric environment and human health. Researchers have shown that gasoline engines, especially Gasoline Direct Injection (GDI) engines, tend to emit large amounts of small size particles compared to Port Fuel Injection (PFI) gasoline engines and diesel engines fitted with Diesel Particulate Filters (DPFs). As a result, the particle number emissions of GDI engines will be restricted by the EU6 legislation. The particulate emission level of GDI engines means that they would face some challenges in meeting the EU6 requirement. This thesis undertakes research in the following area. Firstly, the filtration efficiencies of glass fibre filters were quantified using a Cambustion Differential Mobility Spectrometer 500 (DMS500) to see if all of the particles from the sampled gas can be collected by the filters. Secondly, various valve timings and different injection modes such as double injection with a second injection after compression, single early injection and split early injection were implemented to measure the Particulate Matter (PM) emissions and combustion characteristics of a GDI engine under warm-up operating conditions. Thirdly, the techniques for removing volatile particles were investigated using a catalytic Volatile Particle Remover (VPR) and an Evaporation Tube (ET) with hot air dilution under various test conditions. The results show that for the glass fibre filters tested here, the transmission efficiencies of the particles are very low, indicating that PM sampling using fibre filters is an effective method of studying the particulate emissions from the engine. Particle number emissions using double injection with injection after compression were much higher than those with single injection during the intake stroke. Under 1200 rpm, 110 Nm cold engine operation, no reduction effect on PM emissions was shown by using split intake injection to further facilitate homogeneous mixture formation compared with single intake injection. Valve timings showed moderate effects on particulate emissions. Properly adjusted timing for exhaust valve closure led to reduced particulate emissions by a factor of about 2 and the combustion characteristics were not adversely affected much. The VPR temperature and exhaust residence time did not show much effect on the catalytic VPR performance once the mass flow rate of exhaust was above 0.09 g/s. Generally, the transmission efficiencies of the VPR follow the trends of the scaled PMP counting efficiency specification. Hot air dilution is effective in reducing the small size particles. At 23 nm, the transmission efficiencies are within the error range of the PMP specification. The catalytic VPR and the Evaporation Tube were all found to be effective in reducing the particle number of small size (nucleation mode) particles. Both systems have some particle loss mainly due to the physical effects of diffusion and thermophoresis. Until now, GDI engines have not been optimised for reducing particulate emissions as the focus has been on gaseous emissions and fuel economy. With careful re-optimisation of the catalyst light-off and engine calibration (especially for transients) then there is scope for GDI engines to meet forthcoming emissions legislation.

A thesis submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. / Increasing global energy demand as well as air quality concerns have in recent years led to the search for alternative clean fuels to replace fossil fuels. One such alternative is the blending of petrol (gasoline) with ethanol, which has numerous advantages such as ethanol’s ability to act as oxygenate thus reducing the carbon monoxide emissions from the exhaust of internal combustion engines of vehicles. However, the hygroscopic nature of ethanol is a major concern in obtaining a perfectly homogenized petrol-ethanol fuel. This problem has led to the study of ways of homogenizing the petrol-ethanol mixtures. Therefore, this thesis aimed at enhancing the homogenization of petrol-ethanol mixture. Ethanol concentration in ethanol-water mixture plays a key role in enhancing the homogenization of the fuel, thus the bioethanol employed in this study was dehydrated with silica gel using ultrasonication-enhanced adsorption. Afterwards, the dehydrated ethanol was used in studying the homogenization of the fuel blend. Water removal from the bioethanol using ultrasonication-enhanced adsorption shows a 28% increase when compared to the water removal using magnetic-stirring-enhanced adsorption, During ultrasonication-enhanced adsorption, the estimated adsorption enthalpy was – 1 592.82 J/mol (exothermic) and the entropy was -5.44 J/ K mol, indicating a non-ordered loading of water molecules in the adsorption site. In addition, a modified pseudo second order kinetic model given by was proposed for the ultrasonication-enhanced adsorption process. Effect of temperature during ultrasonication-enhanced adsorption was found to be directly proportional to the amplitude and the pulse rate. However, increase in the amplitudes at lower pulse rates resulted in better cavitation, and hence better adsorption. Furthermore, during phase behavior of ethanol-petrol blend, volume fractions of ethanol and petrol were studied with respect to t the depth within the storage container to confirm homogenization of the blend and time of storage. The binodal curve of the ternary diagram shows an increase of homogeneous region indicating an improved interaction between water and petrol. Therefore, the interesting results regarding the homogenization of the fuel blends resulted from using ultrasonication-enhanced blending were very promising, and could be a platform upon which further research efforts could be built on. The concentration distribution in the reactor showed proof of cavitation formation since in both directions, the variation of concentration with both time and distance was found to be oscillatory. On comparing the profiles in both directions, the concentration gradient, diffusion flux, and energy and diffusion rates were found to be higher in the vertical direction compared to the horizontal direction. It was therefore concluded that ultrasonication creates cavitation in the mixture which enhances mass transfer and mixing of ethanol and petrol. The horizontal direction was found to be the diffusion rate limiting step which proposed that the blender should have a larger height to diameter ratio. It is however recommended that further studies be done on the rate-limiting step so as to have actual dimensions of the reactor. Testing of the blended fuel in internal combustion engine showed an optimal performance of this fuel at 60 % volume ethanol content with higher fuel power. The results of fuel consumption and emissions (such as CO2 and CO) trends confirm various reports in literature on the performance of ethanol/petrol blended fuel.

A presente pesquisa trata de calibrar, a ferir e utiliza em laboratório um detetor de fotoionização (método de campo) proveitoso na caracterização de áreas contaminadas por vazamentos de gasolina; comparar as informações do detetor com os resultados obtidos com balança analítica; comparar os resultados de fluxo de massa medido com o previsto pelo modelo JURY contido no software APIDSS 2.0 O detetor mostrou-se útil como primeira ferramenta na caracterização de áreas contaminadas recentemente (até 100 dias). Para áreas onde a contaminação esteja em uma fase mais avançada o aparelho mostrou-se útil na detecção de alterações; os modelos desenvolvidos com base nos resultados obtido com balança e com o detetor de fotoinização apresentaram resultados satisfatórios para o período de ensaio; as comparações realizadas com o software APIDSS 2.0 serviram para a confirmação dos resultados obtidos. / The present research aims at gauging, checking and using a photoionization detector (field method) in laboratory advantageous to the characterization of areas polluted by leaks of gasoline; to compare the information of the detector with the results obtained in the analytic scale; to compare the results of the measured mass flow with those foreseen by the model JURY of the software APIDSS 2.0. The detector showed to be useful as a first tool in the characterization of areas recently contaminated (up to 100 days). To areas where the contamination is in a more advanced phase, the apparel showed to be useful in the detection of alterations; the development of models based on the results obtained with a scale and with the photoionization detector presented satisfactory results for the rehearsal period; the comparisons accomplished by the software APIDSS 2.0 confirmed the obtained results.

A gasolina que é comercializada no Brasil sofre a adição de álcool etílico anidro, também conhecido como etanol, em frações que podem variar de 20 a 24%. O uso desse composto adicionado à gasolina serve para aumentar a octanagem e reduzir a emissão de monóxido de carbono para a atmosfera. Esta pesquisa teve como objetivo verificar e comparar a relação existente entre a espessura encontrada nos poços de monitoramento e no meio poroso, tanto para gasolina pura quanto para a gasolina com 20 % de etanol (E-20), após a simulação de um vazamento em subsuperfície, utilizando colunas de areia em laboratório, visto que esta relação é de extrema importância para o cálculo preciso do volume do contaminante a ser bombeado, no caso de vazamentos de combustível. As colunas foram confeccionadas utilizando-se 2 tubos de acrílico transparente com 100 cm de comprimento e 23cm de diâmetro interno que foram fixados a uma base de acrílico transparente com 0,60 cm de espessura. Um tubo de PVC branco com 100 cm de comprimento e 3,50 cm de diâmetro interno, com ranhuras horizontais de 0,10 cm de largura foi cortado longitudinalmente e cada metade foi fixada nas paredes dos tubos de acrílico para representar o poço de monitoramento. As colunas foram preenchidas com areia de granulometria muito fina (0,088 mm). Os resultados obtidos mostraram que a espessura aparente da gasolina pura foi 2,6 vezes superior à espessura real da fase livre e a espessura aparente da E-20 foi 0,6 vezes superior a espessura real da fase livre. A tensão interfacial da E-20 foi 67% inferior à tensão interfacial da gasolina pura, possibilitando que uma quantidade maior de poros fosse acessada, refletindo em valores de saturação total 54% superiores ao da gasolina pura. Parte da fase gasolina da E-20 deve ter sido depositada como glóbulos desconectados nos poros caracterizando uma saturação residual, devido ao particionamento preferencial do etanol para a água subterrânea e o conseqüente enriquecimento em gasolina na fase gasolina. Esta saturação residual deve ter dificultado o fluxo e consequentemente reduzido a permeabilidade relativa para a fase aquosa, causando o retardamento da chegada da E-20 no poço de monitoramento. / Commercial gasolines sold in Brazil are added ethanol from 20 to 24% by volume. This oxygenated additive raises the octane level and reduces carbon monoxide emissions to the atmosphere. Underground storage tanks and gasoline lines are a major potential problem for soil and groundwater contamination. Since ethanol has cosolvent properties that alter the behavior of phase distribution in subsurface, this research compares the existing relationship of apparent and actual thicknesses of free phase gasoline and E-20 (80% gasoline and 20% ethanol by volume) found respectively in monitoring wells and medium porous material using two column experiments. Two cylindrical plexiglass columns were used (100 cm long and 23 cm in internal diameter). The bottom ends were capped using two 35 x 35 cm plexiglass plates. A spout-like connection of PVC was fitted near the base of each column as an inlet/outlet end. One well screen (100 cm long and 3,5 cm in internal diameter) in white PVC was cut in half lengthwise and attached to the inside walls of columns to be used as monitoring wells. The columns were then filled with 0,088 mm grain size sand (very fine).The apparent thickness of pure gasoline was 2.6 times thicker than the actual thickness in the aquifer while it was only 0.6 thicker for E-20. The interfacial tension of E-20 is 67% lower than the that of the pure gasoline, allowing a larger quantity of pores to be accessed, which was reflected in an oil saturation 54% higher than that of pure gasoline. Part of the gasoline phase of E-20 must have been precipitated as disconnected doplets within the pores creating a residual saturation and consequently reducing the permeability with respect to the aqueous phase delaying the breakthrough of E-20 in the monitoring well.

O principal objetivo deste estudo foi estimar um modelo analítico que explicasse a relação entre os mercados de etanol hidratado e de gasolina C, incluindo, também, efeitos de outros fatores, como o preço de açúcar nos mercados doméstico e internacional e a renda. Por sua capacidade de captar efeitos dinâmicos, utilizou-se o Modelo de Auto-Regressão Vetorial com Correção de Erro VEC. A análise das funções de demanda dos combustíveis mostrou que o consumo de etanol responde de forma expressiva a variações contemporâneas no próprio preço e no preço da gasolina, produto substituto nesse segmento de mercado. Já com relação à demanda de gasolina C, verifica-se que suas vendas respondem menos tanto ao próprio preço como ao do etanol. De maneira geral, a análise das funções de respostas a impulsos permite verificar que os efeitos dos choques não antecipados sobre as variáveis do modelo são de curta duração, desaparecendo, na maior parte das vezes, após o terceiro mês. A renda mostrou-se importante variável explicativa tanto para o consumo de etanol hidratado quanto de gasolina C. Um choque exógeno no preço internacional de açúcar resulta em um efeito elevado no preço do produto no mercado doméstico. De outro lado, um choque no preço doméstico de açúcar tem impacto significativo no preço do etanol anidro (elasticidade acumulada de 0,57), sendo que esses produtos concorrem pela mesma matéria-prima na indústria. Como este preço, por sua vez, representa uma variável proxy para o preço do etanol hidratado no segmento produtor, pode-se afirmar que 60% das variações de preço neste segmento são transmitidas ao preço do combustível no mercado varejista. A análise da decomposição das variâncias do erro de previsão evidenciou que a variável que tem maior impacto sobre o preço do etanol hidratado combustível no varejo é o preço do etanol anidro (proxy do preço do etanol hidratado no segmento produtor). Os principais determinantes da venda de etanol hidratado no varejo são a renda, o preço do etanol hidratado combustível no varejo e o do etanol anidro no segmento produtor. Já com relação às vendas de gasolina C no varejo, o principal determinante é a renda, enquanto as demais variáveis pouco contribuem para explicar suas variações. Finalmente, a análise da previsão dentro da amostra para o preço do etanol hidratado no segmento varejista sugere que este preço assumiria um patamar mais elevado, não fosse a atual política de controle de preços dos derivados do petróleo adotada no Brasil. Nas duas últimas safras do período estudado, o preço praticado está, em média, 23,7% abaixo do previsto pelo modelo estimado neste estudo. Vale ressaltar que o modelo estimado capta o padrão médio do comportamento da série, dados os fundamentos de mercado considerados, os quais explicam pelo menos 76% das variações do preço do etanol hidratado no varejo. / The main objective of this study was to estimate an analytical model to explain the relationship between the markets of hydrated ethanol fuel and gasoline C, including effects of other factors such as sugar prices in the domestic and the International markets and income. Because of its ability to capture dynamic effects, the Auto-regression Model with Vector Error Correction VEC was used. Analysis of the demand function showed ethanol fuel consumption significantly responds to contemporary variations of its price and gasoline price, replacement product in this market segment. Regarding the demand for gasoline C, it appears that sales respond less to its price and ethanol price. In general, analysis of the response functions to impulses allows to verify that the effect of unanticipated shocks on model variables are of short duration and disappears, in most cases, after three months. Income proved to be important in ethanol and gasoline C consumption. An exogenous shock in the international sugar price results in a large effect on the product price in the domestic market. Therefore, a shock in the domestic sugar price has a significant impact on the anhydrous ethanol price (accumulated elasticity of 0.57), once these products compete for raw material in the sector. Given that this price represents a proxy variable for the hydrous ethanol price for the industry, 60% of the price in this segment is transmitted to the fuel price in the retail market. A decomposition analysis of forecast error variance showed that the variable with greater impact on the hydrated ethanol fuel price in the retail market is the anhydrous ethanol price (proxy of the hydrated ethanol price in the industry). The main determinants of hydrated ethanol sales in the retail market are the income, the hydrated ethanol fuel price and the anhydrous ethanol fuel price in the industry. Gasoline C sales in the retail market are mostly determined by income, while the other variables contribute little to explain its variations. Finally, the analysis of the in-sample forecast for the hydrated ethanol price in the retail market suggests that this price would assume a higher level, if there were not the current policy of controlling prices of petroleum derivatives adopted in Brazil. In the last two seasons of the analysis, the price was on average 23.7% below that predicted by the model estimated in this study. It is noteworthy that the estimated model captures the average standard of the series, given the market fundaments considered, which explain at least 76% of the variations in the hydrated ethanol price in the retail market.

Dados da CETESB revelam que a comercialização de derivados do petróleo na Cidade de São Paulo, em especial a gasolina, tem gerado um número significativo de casos de vazamento de combustíveis, causados pela falta de manutenção de equipamentos, deterioração de tanques e tubulações, e falhas operacionais decorrentes do despreparo dos profissionais que atuam nesses estabelecimentos. A área urbana da Cidade de São Paulo está, em grande parte, edificada sobre os depósitos sedimentares de idade cenozóica da Bacia de São Paulo, onde a Formação Resende apresenta distribuição generalizada e compreende mais de 80% do preenchimento sedimentar. Argilominerais esmectíticos predominam na fração fina dos sedimentos da Formação Resende e a possível interação deles com hidrocarbonetos foi o objeto de investigação dessa dissertação. Nesse sentido, o presente trabalho teve por objetivo o estudo laboratorial da interação, por meio de sorção e desorção, de hidrocarbonetos encontrados comumente, como gasolina comercial, benzeno, tolueno e xilenos, com argilominerais esmectíticos da Formação Resende, no contexto da zona vadosa, na Bacia de São Paulo. Para tanto, três amostras argilosas da Formação Resende foram coletadas na zona leste da Cidade de São Paulo, tendo sido efetuada sua caracterização granulométrica, mineralógica (via DRX) e química (análise de CNH). Em testes de inchamento com gasolina e orgânicos puros, nenhuma das amostras apresentou absorção dos compostos. Os testes de adsorção e desorção procuraram caracterizar a interação dos hidrocarbonetos com a rocha total, fração silte e fração argila dos sedimentos, variandose o tempo de exposição (30 minutos a 32 dias) das amostras aos contaminantes. Os resultados obtidos permitiram reconhecer que a adsorção é o mecanismo de sorbância predominante. A variação das concentrações ao longo do tempo evidenciou que a interação amostra/contaminante é dinâmica ao longo do tempo de contato, havendo momentos distintos de sorbância, com variação significativa e bem representada dos teores adsorvidos, e que essas fases ocorrem independentemente da granulometria ou mineralogia das amostras, sendo: os sedimentos argilosos adsorvem, na média, as mesmas ordens de grandeza de cada contaminante, o que é devido a sua homogeneidade granulométrica e mineralógica em pequena escala; dentro da ordem de grandeza de adsorção de cada contaminante, as diferenças composicionais (granulométricas e mineralógicas) entre as amostras de sedimentos argilosos alteram os teores dos compostos em função da proporção de argila na rocha total, de forma que quanto mais contaminante será adsorvido, e do tipo de argilomineral predominante na fração argila, sendo a maior proporção de esmectita um fator favorável. / Data from the Environmental Agency of São Paulo State (Company of Environmental Sanitation Technology CETESB) reveal that the commercialization of petroleum products in the city, especially gasoline, has generated a high number of fuel leakages and contaminated areas. The urban area of São Paulo city is underlain by the Cenozoic sedimentary deposits of the São Paulo Basin. In this basin, smectiterich mudstone of the Resende Formation is widely distributed and comprises over 80% of the sedimentary filling. The aim of this study was to investigate the interaction of smectiterich mudstone of the Resende Formation with hydrocarbons in the São Paulo Basin. The sorption and desorption of common hydrocarbons, such as gasoline, benzene, toluene and xylene, by the smectitic clay minerals, in the context of vadose zone, were investigated in laboratory. Three representative subsurface samples of mudstone were collected in the east part of São Paulo city and characterized by granulometry, Xray analysis and CHN determination. Swelling test with gasoline and pure organic compounds does not promote absorption in the three studied samples. Adsorption and desorption tests were employed to characterize the interaction between the hydrocarbons and the wholerock, silt and clayfractions of the mudstone in different time intervals of contact (30 minutes to 32 days). The results revealed the adsorption as the main interaction mechanism. Variations in concentration of contaminants in different time intervals of contact indicated a dynamic interaction between sample and hydrocarbons, existing phases with distinct amounts of organics adsorbed (or retained after desorption). Theses phases are controlled by the time of contact and are independent of samples granulometry and mineralogy, and contaminant type. The three mudstone samples adsorb about the same amount of each contaminant, which is due to their overall granulometric and mineralogic homogeneity. Nevertheless, minor and local changes in sample composition (granulometry and mineralogy) cause a small scale variation in the amounts of each contaminant adsorbed as a result either of clay proportion in the whole-rock and of the predominant clay mineral in the clay fraction.

Time-delay feed-forward Artificial Neural Networks are examined for gasoline engine cylinder pressure reconstruction using both measured crank kinematics obtained from a shaft encoder, and measured engine cylinder block vibrations obtained from a production knock sensor. Initially, the study focuses on the information content associated with measured data, which is considered to be of equal importance to the particular network architecture and the training methodology. Several hypotheses are constructed, which when tested, reveal the influence of the data information content on the reconstruction potential and limitations. These hypotheses are tested on real data from a 3-cylinder (DISI) engine. Three distinct ideas emerge through this testing process, which are combined to produce a single pressure reconstruction methodology. Reconstruction results obtained via this methodology, applied to crank kinematics associated with steady-state engine operation, show a marked improvement over previously published reconstruction accuracy. Moreover, in steady-state engine operation, the application of this methodology to acceleration measurements of cylinder block vibration, obtained from a knock sensor, show very significant improvements over previous attempts. But the direct application of this same reconstruction methodology to transient engine operation, proves to be problematic. However, a novel generalisation of the approach in the form of a time-dependent feed-forward neural network is proposed and the required adaptation made to the use of the Levenberg-Marquardt training algorithm. This time-dependent approach has been tested under limited transient conditions and shown in the thesis to give good results, therefore offering considerable potential for use with real engine operation. Overall, the thesis shows that by careful processing of measured engine data, standard neural network architectures and standard training algorithms can be used to reconstruct engine cylinder pressure.