<p>This study was motivated by how to use statistical tool to identify the candidacy wells for CO2 Capture and Sequestration based on the idea of using Artificial Neural Networks to predict the leakage index of a well. A Combined Neural-Genetic Algorithm was introduced to avoid BP neural network getting a local minimum because Genetic Algorithm simulates the survival of the fittest among individuals over consecutive generation. Based on the algorithm, 1356 lines of C code were composed using Microsoft Visual Studio 2010. The Combined Neural-Genetic Algorithm developed in this thesis is able to handle large size of data sample with at least 10 factors. Several parameters were considered as factors that may have an effect on the performance of Combined Neural-Genetic Algorithm, including the population size, max epoch, error goal, probability of crossover, probability of mutation, number of neurons in hidden layer, number of factors and size of data sample. The accuracy of the BP neural network and the CPU time are the two major parameters to evaluate the performance of the Combined Neural-Genetic Algorithm. A sensitivity analysis was performed to identify the effect these factor have on the performance. Based on the result of the sensitivity analysis, some recommendations are provided about initializing these factors.

Gas flow behavior from a spherical source is explored by using linear and nonlinear models, not only in terms of pressure but also in terms of flux. The approach considers dimensionless parameters scaling both radius and time. Specific observations are made for large, moderate, and small time conditions. At large time, the nonlinear model becomes a linear ordinary differential equation with pressure solution independent of the material. However, for moderate and small scaled times this is not the case. The nonlinear model must be solved by using either linear approximations, semi-analytical, or numerical procedures. This model is nonlinear in the primary variable (pressure). However, appropriate mathematical manipulations allow one to change the nonlinearity into a single coefficient, depending on pressure. Focusing on the effects of this coefficient, the nonlinear solution can be confined between two linear solutions obtained by using atmospheric and boundary pressures. Appendix A is an exploration of the errors arising between the nonlinear solution and these two solutions. In Appendix B, a nonlinear model is used to find solutions for large, moderate, and small times. For large time, the case corresponds to the steady state case, and coincides with the solution presented in Appendix A. For moderate and small times the quasi-analytical approximation and the asymptotic solutions of linear and quadratic normalizations of pressure are presented. In Appendix C, simulations of gas flows in linear and nonlinear situations are made. The problem is to determine the change of air pressure in a tank when it is connected to a spherical cavity embedded in a porous medium. These changes in pressure occur when the air moves through the porous media, either for gas extraction or air injection. Both linear and nonlinear analyses require calculations of the pressure and the mass in the tank when the initial and boundary conditions change with time. For each case, gas extraction or air injection, the differences between the linear and the nonlinear models are examined to determine the suitability of the linear model.

A model consisting of the Hamaker-Lifshitz theory is used to calculate the spreading transition of alkanes at the water-gas interface. In the current work, the theory is extended to include effects of the temperature and pressure by estimating dielectric polarizability of the vapor and liquid as a function of density through the Clausius-Mosotti equation. The wetting transition temperature of alkane at the water-gas interface increases with increasing of carbon number. The theory is validated by measuring the contact angle of alkane at the water-gas interface by using an interference microscope. Spreading coefficients calculated from contact angles agree with the calculated values. The Hamaker constant calculation, measured contact angle, and measured spreading coefficient show that alkanes approach spreading as the NaCl concentration in water approaches saturation. The theory of this study is also used to predict the flow of sulfur containing sour natural gas in a reservoir, and the calculations are supported by the experimental results.

This project culminated in the development of a new flow method and device for measuring the water contents of high pressure gases. This new flow method uses both an electrical resistance sensor and/or a chemical desiccant method to measure the water contents of methane, ethane and methane + carbon dioxide and ethane + carbon dioxide gas mixtures from 3.4 MPa to 110 MPa over a temperature range from 310 to 477 K. The resulting measurements have reduced uncertainty in the binary experimental results to between 2 to 7 percent, and give ternary results with an uncertainty between 5 and 14 percent. The new experimental data are modeled using a hybrid method which combines the Peng-Robinson equation of state to calculate vapor phase fugacity coefficients with a highly accurate equation of state to calculate the fugacity of water. Gas solubilities in the aqueous phase are calculated using a Henry's law coefficient, while aqueous fugacities are calculated using the NIST/ASME equation of state for water, while the effects of salts are incorporated using Pitzer correlations for the activity of brines. The new model is able to predict ternary phase equilibria using interaction parameters fit to binary data. This allows predictive phase behavior calculations to be made for multiple components. The related system methanol + alkanes, where the alkane was a member of the homologous series propane to decane was modeled using the Statistical Associating Fluid Theory (SAFT) equation of state for both liquid-vapor (VLE) and liquid-liquid equilibria (LLE). It was shown that the PC and CK-SAFT equations of state were capable of representing the phase behavior to within a few percent (generally 1--4%) of the experimental data using binary interaction parameters that were weak linear functions of temperature and alkane molecular weight. The binary interaction parameters were fit to the VLE data and then applied to the LLE data with excellent results for the methanol + alkane systems from 1 to 150 MPa. For alkanes longer than octane, systematic deviation was observed.

Asphaltene precipitation and deposition can occur at different stages during petroleum production causing reservoir formation damage and plugging of pipeline and production equipment. Even though asphaltene deposition is a serious production hazard, deposition appears to occur only if precipitation is present. The main motivation of this work is to develop a general model for asphaltene precipitation and to understand the contribution of the surface material to asphaltene deposition. This dissertation presents a study of the effects of temperature, pressure, and composition on asphaltene phase separation prediction using the Statistical Associating Fluid Theory (SAFT) equation of state (EOS) and application of the theory to field cases. Furthermore, a molecular theory is used to predict the deposition tendency of asphaltene. Practical understanding of asphaltene precipitation applied to the oil field production is presented in this research project. A challenge overcame in this study was to translate this methodology to industrial application; initially, through the validation of the PC-SAFT EOS model as implemented in commercial computer software. Then, this work shows how SAFT qualitatively predicts different scenarios as actually occur in the field, i.e., the effect of gas injection, specifically, CO2, N2 and normal alkanes; oil based mud contamination and commingling of different live oil streams. These aspects were studied using crude oils from Deepwater Gulf of Mexico and Middle East. Finally, the impact of asphaltene precipitation considerations in a deepwater development project was studied based on experimental measurements of a hydrocarbon fluid when contacted with gas condensate from another zone. The evaluation was performed using multiphase thermal-hydraulic behavior coupled with the asphaltene PC-SAFT thermodynamic model. The PC-SAFT EOS adequately predicts the onset of asphaltene precipitation in all these cases. Simulation results agree with previous experimental reported work. On the depositional aspect a theoretical evaluation of the asphaltene adsorption behavior onto solid surfaces has been made to look for a relationship between the composition of the solution phase and the surface through the application of molecular theory. The asphaltene deposition tendency can be qualitatively described through the conventional Hamaker constant, which represents molecular van der Waals interactions between macroscopic bodies. Results show agreement with experimental observations.

The default assumptions used for interpreting Nuclear Magnetic Resonance (NMR) measurements with reservoir rocks fail for many sandstone and carbonate formations. This study provides quantitative understanding of the mechanisms governing NMR relaxation of formation fluids for two important cases in which default assumptions are not valid. The first is diffusional coupling between micro and macropore, the second is susceptibility-induced magnetic field inhomogeneities. Understanding of governing mechanisms can aid in better estimation of formation properties such as pore size distribution and irreducible water saturation. The assumption of direct correspondence between relaxation time and pore size distribution of a rock fails if fluid in different sized pores is coupled by diffusion. Pore scale simulations of relaxation in coupled micro and macropores are done to analyze the effect of governing parameters such as surface relaxivity, pore geometry and fluid diffusivity. A new coupling parameter (alpha) is introduced which quantifies the extent of coupling by comparing the rate of relaxation in a coupled pore to the rate of diffusional transport. Depending on alpha, the pores can communicate through total, intermediate or decoupled regimes of coupling. This work also develops a new technique for accurate estimation of irreducible saturation, an approach that is applicable in all coupling regimes. The theory is validated for representative cases of pore coupling in sandstone and carbonate formations. Another assumption used in NMR formation evaluation is that the magnetic field distribution in the pores corresponds to the externally applied field. However, strong field inhomogeneities can be induced in presence of paramagnetic minerals such as iron on pore surfaces of sedimentary rocks. A generalized relaxation theory is proposed which identifies three asymptotic relaxation regimes of motionally averaging, localization and free diffusion. The relaxation characteristics of the asymptotic regimes such as T 1/T2 ratio and echo spacing dependence are quantitatively illustrated by random walk simulations and experiments with paramagnetic particles of several sizes. The theory can aid in better interpretation of diffusion measurements in porous media as well as imaging experiments in Magnetic Resonance Imaging (MRI).

A high pressure heat flux calorimeter in isobaric, temperature-ramping mode has been used to measure the solubility of pure methane, pure carbon dioxide and methane-carbon dioxide mixtures. The solubility measurements emphasize a crystallization-like process taking place during hydrate formation and show a striking divergence from Henry's Law, the frequently used calculation procedure, prior to and during hydrate formation. These measurements were further used to determine the enthalpies of solution/dissociation, and entropies change. Moreover, the hydration numbers of these compounds provide some explanations and criteria of stability of the cages of gas hydrates in the host lattice. Finally, a kinetic study confirms the crystallization process of hydrate formation and exhibits a high level of supersaturation prior to hydrate formation and also a high consumption rate during hydrate formation.

This research explores a new method using foams for water production control in an oilfield. Reducing water production during oil production is an important objective impacting the profitability of a mature oilfield. Currently practiced methods using gel or polymer based systems either offer inadequate water flow reduction or suffer problems of proper placement in the field. Because of its properties, foam has the potential for use in water control. In this study, foams stable in presence of flowing water (washout stability) were developed using polymeric surfactants. A screening test was developed to measure the washout resistance of various conventional and polymeric surfactants. Foam from several polymeric surfactants such as triblock F108 and hydrophobically modified HMPA1 exhibited remarkable improvement in washout stability over conventional surfactants. Strong foam that offered a large resistance to flow of water was generated in a two-foot long sand pack with some of these polymeric surfactants. Again, the polymeric surfactants exhibited higher foam washout resistance than the conventional surfactants as predicted by the screening tests. Investigation of surfactant desorption from an air-water interface using bubble shape analysis showed that this improved foam washout resistance was due to almost irreversible adsorption of polymeric surfactants. Collapse of foam from polymeric surfactants at long times in the screening test was determined to be due to hydrodynamic effects and not desorption. Also, foam washout stability with polymeric surfactants in sand pack was found to be limited by air dissolution into flowing water. Scale-up calculations for oilfield geometries showed that foam from F108 can be stable for a long enough time, even with gas dissolution, for the process to be practicable. Foam stable to residual oil, expected in the water producing zones, was created by mixing an anionic surfactant CS-330 with nonionic F108. This is because ionic surfactants produce an electrostatic barrier that prevents entry of oil droplets into the air water interface. Flowing oil, however, produced a stable emulsion with this surfactant combination which offered a large resistance to flow. This was undesirable and was minimized by a brine flush to remove surfactant from the aqueous phase of the foam region before contact with flowing oil.

Oil recovery by water flooding is usually not effective because of capillary forces in fractured, oil-wet carbonate formations. Sodium carbonate/surfactant solution was used to enhance spontaneous imbibition between the fractures and the matrix by both wettability alteration and ultra-low interfacial tensions. Carbonate formations are usually oil-wet because the mineral-brine and oil-brine interfaces have zeta potentials of opposite sign. The resulting electrostatic attraction destabilizes the water film between the mineral and crude oil. The zeta potential of calcite can be reversed to be negative even at neutral pH, with dilute solutions of sodium carbonate and bicarbonate. Carbonate ion is a potential determining ion for carbonate formations. Thus sodium carbonate can promote water-wetness. Carbonate ion also sequesters calcium from the brine because of the small solubility product. The negative zeta potential and low calcium concentration greatly reduce the adsorption of anionic surfactants on the surface of carbonates. Another important effect of sodium carbonate is to generate natural soap in situ by saponifying organic acids in the crude oil. Optimal salinity was found to depend only on the soap-to-surfactant ratio. Below optimal salinity, sometimes a thin layer accumulated between the lower-phase microemulsion and excess oil. The IFT of excess oil with the equilibrated lower-phase microemulsion was high. However, ultra-low interfacial tension was observed when material from the thin layer was dispersed in the lower-phase or added to it. The existence of this thin layer made wide ultra-low interfacial tension possible. The alteration of wettability is graphically illustrated by observation with a polished marble plate. After aging in crude oil, the plate is strongly oil-wet in brine. When the brine is replaced with a sodium carbonate/surfactant solution, wettability can be altered to water-wetness to intermediate wetness. When sodium carbonate is the only salt, both drop size and contact angle are found to decrease with salinity to an equilibrium value. But when sodium carbonate concentration is fixed at one per cent, and sodium chloride is used to change ionic strength, contact angle is not found to change much with salinity, but drop size goes through a minimum at a salinity corresponding to the optimal salinity at the experimental condition. Oil in a narrow gap between two parallel plates remains in place when submerged in brine because of capillary forces. However, this oil is displaced by buoyancy when the brine is replaced with a sodium carbonate/surfactant solution. Displacement rate is found to be dependent on electrolyte concentration. No spontaneous imbibition occurred when a partially oil saturated reservoir core sample was placed in formation brine. Oil was spontaneously displaced when the brine was replaced with a sodium carbonate/surfactant solution. The recovery rate was found to scale with gravity drainage, not capillary imbibition.

A 90 MHz NMR Spectrometer equipped with a high pressure probe was used to study relationship between NMR relaxation time and temperature, viscosity, diffusivity and gas/oil ratio of methane-hydrocarbon mixtures. This research project involves three parts: (1) modifications of the existing NMR apparatus. (2) Measurements of relaxation times and diffusion coefficients of methane-hydrocarbon mixtures. (3) Development of generalized correlations between transport properties and temperature and relaxation times. The NMR apparatus was modified in order to make elevated temperature and pressure measurements. The modifications included calibration of pressure transducers, addition of temperature measuring devices, connection to the high pressure sample probe of a sapphire sample cell and leak detection of the system. After the modifications, the apparatus was capable of measurements from 20 to 60&deg;C at pressure up to 6000 psia. NMR relaxation measurements of three mixtures, methane-n-hexane, methane-n-decane and methane-n-hexadecane, were made. The log mean relaxation times were plotted against viscosity/temperature and it was found that unlike stock tank oils, they do not depend linearly on viscosity/temperature on a log-log scale. Each of the mixtures forms a different curve on the plot of relaxation time vs. viscosity/temperature. Diffusivity measurements were also made for these three mixtures, as well as pure hexane, decane and hexadecane. The log mean diffusion coefficients were calculated. The relationship between diffusion coefficients and relaxation times were studied, and it was found that diffusion coefficients depend linearly on T1 for pure hydrocarbons, but the dependence does not hold for methane-hydrocarbon mixtures. Correlations between transport properties and NMR relaxation times were developed. First, a relaxation time mixing rule was developed by studying the theory of NMR relaxation mechanism. From the mixing rule, it was found that departure of relaxation times of methane-n-alkane mixtures from linear correlation on a log-log scale can be correlated with proton fractions of methane, which can be expressed as gas/oil ratio. Thus, correlation between relaxation time, viscosity/temperature and gas/oil ratio was developed. Correlation between relaxation time, diffusivity and gas/oil ratio was also developed. From these correlations, viscosity and gas/oil ratio can be estimated just from NMR relaxation time and diffusion coefficient.