Highly accurate data for density measurements are required for engineering calculations as well for developing equations of state (EOS) for use in the custody transfer of natural gas through pipelines. The widely used present industry standard, the AGA8-DC92 EOS, was developed against a database of reference quality natural gas mixtures with compositions containing less than 0.2 mole percent of the heavier C6+ fraction. With the advances in technology in the late nineties, it is possible to produce gas from deep and ultra-deepwater of the Gulf of Mexico where the pressures and temperatures encountered are much higher. Produced gas mixtures have compositions containing higher percentages of the C6+ fraction. As AGA8-DC92 is a statistical fit equation developed for one set of conditions, time has come to evaluate its performance to assess whether it is still viable for gas custody transfer with a new set of conditions encountered. A highly accurate, high pressure and temperature, compact single sinker magnetic suspension densimeter has been used first to determine densities of pure component’s densities for which very reliable data are available. After validating its performance, the densities of four light natural gas mixtures, containing no C6+ fraction and two heavy gas mixtures containing more than 0.2 mole percent of the heavier C6+ fraction, were measured. The light mixtures were measured in the temperature range of 250 to 450 K and in the pressure range of 10 to 150 Mpa (1450 to 21,750 psi); the heavy mixtures were measured in the range of 270 to 340 K and in the pressure range of 3.45 to 34.45 MPa (500 to 5,000 psi). Out of those, the data of only two light natural gas mixtures have been presented in the dissertation. Data on two heavy mixtures have not been published due to reasons of confidentiality. Measured densities of light mixtures, not containing the C6+ fraction show less than expected relative deviations from the AGA8-DC92 EOS predictions except at low temperature. The deviation with the recently developed GERG02 EOS was more pronounced. A force transmission error analysis and uncertainty analysis was carried out. The total uncertainty was calculated to be 0.105 %. The data measured as a part of this research should be used as reference quality data either to modify the parameters of AGA8-DC92 EOS or develop a more reliable equation of state with wider ranges of pressure and temperature.

High-temperature, high-pressure (HTHP) conditions are exemplified in ultra-deep petroleum reservoirs and can be exhibited within diesel engines. Accurate pure component hydrocarbon data is essential in understanding the overall behavior of petroleum and diesel fuel at these conditions. The present study focuses on the HTHP properties of toluene since this hydrocarbon is frequently used to increase the octane rating of gasoline and toluene occurs naturally in crude oil. In this thesis experimental densities and viscosity are presented to 535 K and 300 MPa extending the database of toluene viscosity data to higher temperature than previous studies. The data is correlated to a Tait-like equation and a Padѐ approximate in conjunction with a single mapping of the isotherms. Free-volume theory and a superposition of the viscosity in relation to the Leonnard-Jones repulsive force are both used to model the toluene viscosity data. It was found that the data are in good agreement with the available literature data.

[pt] No âmbito das cervejarias artesanais, foi recentemente inventado um medidor de densidade de líquidos que opera por um princípio incomum: o do deslocamento do centro de carena. Esse medidor obtém suas medições a partir da sua própria inclinação enquanto está flutuando, mas sua implementação original converte as medidas do acelerômetro em medidas de massa específica por uma regressão polinomial. Contudo, ele não faz correções de temperatura, de forma que a influência dessa grandeza é desconsiderada na regressão. Adicionalmente, o medidor não indica qual a sua incerteza de medição. Esta dissertação teve como objetivos criar um modelo matemático do fenômeno, que não foi localizado na bibliografia existente; utilizar o modelo para a obtenção de uma estimativa da incerteza de medição, comparando as metodologias de incerteza do GUM e a que utiliza o método de Monte Carlo do Suplemento 1 e utilizando a segunda abordagem para validar a primeira; executar experimentos com um protótipo de um medidor desse tipo, comparando os resultados com um densímetro de laboratório; e realizar a caracterização metrológica do medidor. Todos esses objetivos foram cumpridos, sendo que a caracterização incluiu: sugestões de procedimentos de calibração e de medição; os resultados do experimento, incluindo a distribuição esperada na saída, com média Peso de um objeto – para um sólido rígido = 1,0500 g/cm(3) e incerteza expandida máxima de U95 por cento(p) = U95 por cento(1,0000) = 0,0028 g/cm(3) (fator de abrangência k = 1,96) no intervalo de medição entre 1,0000 g/cm(3) e 1,1000 g/cm(3); equações para a estimativa da incerteza de medidores desse tipo; a estimação de uma curva de incerteza para a faixa de calibração, segundo as medições de calibração; as contribuições de cada grandeza de entrada sobre a incerteza estimada de saída e algumas sugestões de como o medidor poderia ser modificado para melhorar o resultado. / [en] In the context of craft breweries, a recently invented liquid density meter works by leveraging an unusual principle: the displacement of the center of buoyancy. This meter obtains its measurements from its own tilt while it is floating, but its original implementation converts the accelerometer measurements into density measurements with a polynomial regression. However, it doesn t make temperature corrections, so that the influence of this quantity is disregarded in the regression. Additionally, the meter does not indicate its measurement uncertainty. The objective of this dissertation was to create a mathematical model of the phenomenon, which was not found in the existing bibliography; use the model to obtain an estimate of the measurement uncertainty, comparing the uncertainty methodologies of the GUM and the one that uses the Monte Carlo method of its Supplement 1 and using the second approach to validate the first; perform experiments with a prototype of such a meter, comparing the results with a laboratory densimeter; and perform the metrological characterization of the meter. All these objectives were met, and the characterization included: a suggestion of calibration and measurement procedures; the results of the experiment, including the expected output distribution, with mean of Weight of an object – for a rigid solid= 1.0500 g/cm(3) and maximum expanded uncertainty U95 percent(p) = U95 percent(1.0000) = 0.0028 g/cm(3) (k = 1.96 coverage factor) in the measurement range between 1.0000 g/cm(3) and 1.1000 g/cm(3); equations for estimating the uncertainty of this type of meter; the estimation of an uncertainty curve for the calibration range, according to the calibration measurements; the contributions of each input quantity to the estimated output uncertainty and some suggestions on how the meter could be modified to improve the result.