Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.

Finding and developing oil and gas resources requires accurate geological information with which to formulate strategies for exploration and exploitation ventures. When data are scarce, statistical procedures are sometimes substituted to compensate for the lack of information about reservoir properties. The most modern methods incorporate geostatistics. Even the best geostatistical methods yield results with varying degrees of uncertainty in their solutions. Geological information is, by its nature, spatially limited and the geoscientist is handicapped in determining appropriate values for various geological parameters that affect the final reservoir model (Massonnat, 1999). This study focuses on reservoir models that depend on geostatistical methods. This is accomplished by quantifying the uncertainty in outcome of reservoir simulations as six different geological variables are changed during a succession of reservoir simulations. In this study, variations in total fluid produced are examined by numerical modeling. Causes of uncertainty in outcomes of the model runs are examined by changing one of six geological parameters for each run. The six geological parameters tested for their impact on reservoir performances include the following: 1) variogram range used to krig thickness layers, 2) morphology around well 14, 3) shelf edge orientation, 4) bathymetry ranges attributed for each facies, 5) variogram range used to simulate facies distribution, 6) extension of the erosion at top of the reservoir. The parameters were assigned values that varied from a minimum to a maximum quantity, determined from petrophysical and core analysis. After simulation runs had been completed, a realistic, 3-dimensional reservoir model was developed that revealed a range of reservoir production data. The parameters that had the most impact on reservoir performance were: 1) the amount of rock eroded at the top of the reservoir zone and 2) the bathymetry assigned to the reservoir facies. This study demonstrates how interaction between geological parameters influence reservoir fluid production, how variations in those parameters influence uncertainties in reservoir simulations, and it highlights the interdependencies between geological variables. The analysis of variance method used to quantify uncertainty in this study was found to be rapid, accurate, and highly satisfactory for this type of study. It is recommended for future applications in the petroleum industry.

Previous experiments - injecting pure CO2 into carbonate cores - showed that the process is a win-win technology, sequestrating CO2 while recovering a significant amount of hitherto unrecoverable natural gas that could help defray the cost of CO2 sequestration. In this thesis, I report my findings on the effect of flue gas ??impurities?? on the displacement of natural gas during CO2 sequestration, and results on unconfined compressive strength (UCS) tests to carbonate samples. In displacement experiments, corefloods were conducted at 1,500 psig and 70??C, in which flue gas was injected into an Austin chalk core containing initially methane. Two types of flue gases were injected: dehydrated flue gas with 13.574 mole% CO2 (Gas A), and treated flue gas (N2, O2 and water removed) with 99.433 mole% CO2 (Gas B). The main results of this study are as follows. First, the dispersion coefficient increases with concentration of ??impurities??. Gas A exhibits the largest dispersion coefficients, 0.18-0.25 cm2/min, compared to 0.13-0.15 cm2/min for Gas B, and 0.15 cm2/min for pure CO2. Second, recovery of methane at breakthrough is relatively high, ranging from 86% OGIP for pure CO2, 74-90% OGIP for Gas B, and 79-81% for Gas A. Lastly, injection of Gas A would sequester the least amount of CO2 as it contains about 80 mole% nitrogen. From the view point of sequestration, Gas A would be least desirable while Gas B appears to be the most desirable as separation cost would probably be cheaper than that for pure CO2 with similar gas recovery. For UCS tests, corefloods were conducted at 1,700 psig and 65??C in such a way that the cell throughput of CO2 simulates near-wellbore throughput. This was achieved through increasing the injection rate and time of injection. Corefloods were followed by porosity measurement and UCS tests. Main results are presented as follows. First, the UCS of the rock was reduced by approximately 30% of its original value as a result of the dissolution process. Second, porosity profiles of rock samples increased up to 2.5% after corefloods. UCS test results indicate that CO2 injection will cause weakening of near-wellbore formation rock.

Finding and developing oil and gas resources requires accurate geological information with which to formulate strategies for exploration and exploitation ventures. When data are scarce, statistical procedures are sometimes substituted to compensate for the lack of information about reservoir properties. The most modern methods incorporate geostatistics. Even the best geostatistical methods yield results with varying degrees of uncertainty in their solutions. Geological information is, by its nature, spatially limited and the geoscientist is handicapped in determining appropriate values for various geological parameters that affect the final reservoir model (Massonnat, 1999). This study focuses on reservoir models that depend on geostatistical methods. This is accomplished by quantifying the uncertainty in outcome of reservoir simulations as six different geological variables are changed during a succession of reservoir simulations. In this study, variations in total fluid produced are examined by numerical modeling. Causes of uncertainty in outcomes of the model runs are examined by changing one of six geological parameters for each run. The six geological parameters tested for their impact on reservoir performances include the following: 1) variogram range used to krig thickness layers, 2) morphology around well 14, 3) shelf edge orientation, 4) bathymetry ranges attributed for each facies, 5) variogram range used to simulate facies distribution, 6) extension of the erosion at top of the reservoir. The parameters were assigned values that varied from a minimum to a maximum quantity, determined from petrophysical and core analysis. After simulation runs had been completed, a realistic, 3-dimensional reservoir model was developed that revealed a range of reservoir production data. The parameters that had the most impact on reservoir performance were: 1) the amount of rock eroded at the top of the reservoir zone and 2) the bathymetry assigned to the reservoir facies. This study demonstrates how interaction between geological parameters influence reservoir fluid production, how variations in those parameters influence uncertainties in reservoir simulations, and it highlights the interdependencies between geological variables. The analysis of variance method used to quantify uncertainty in this study was found to be rapid, accurate, and highly satisfactory for this type of study. It is recommended for future applications in the petroleum industry.

Previous experiments - injecting pure CO2 into carbonate cores - showed that the process is a win-win technology, sequestrating CO2 while recovering a significant amount of hitherto unrecoverable natural gas that could help defray the cost of CO2 sequestration. In this thesis, I report my findings on the effect of flue gas ??impurities?? on the displacement of natural gas during CO2 sequestration, and results on unconfined compressive strength (UCS) tests to carbonate samples. In displacement experiments, corefloods were conducted at 1,500 psig and 70??C, in which flue gas was injected into an Austin chalk core containing initially methane. Two types of flue gases were injected: dehydrated flue gas with 13.574 mole% CO2 (Gas A), and treated flue gas (N2, O2 and water removed) with 99.433 mole% CO2 (Gas B). The main results of this study are as follows. First, the dispersion coefficient increases with concentration of ??impurities??. Gas A exhibits the largest dispersion coefficients, 0.18-0.25 cm2/min, compared to 0.13-0.15 cm2/min for Gas B, and 0.15 cm2/min for pure CO2. Second, recovery of methane at breakthrough is relatively high, ranging from 86% OGIP for pure CO2, 74-90% OGIP for Gas B, and 79-81% for Gas A. Lastly, injection of Gas A would sequester the least amount of CO2 as it contains about 80 mole% nitrogen. From the view point of sequestration, Gas A would be least desirable while Gas B appears to be the most desirable as separation cost would probably be cheaper than that for pure CO2 with similar gas recovery. For UCS tests, corefloods were conducted at 1,700 psig and 65??C in such a way that the cell throughput of CO2 simulates near-wellbore throughput. This was achieved through increasing the injection rate and time of injection. Corefloods were followed by porosity measurement and UCS tests. Main results are presented as follows. First, the UCS of the rock was reduced by approximately 30% of its original value as a result of the dissolution process. Second, porosity profiles of rock samples increased up to 2.5% after corefloods. UCS test results indicate that CO2 injection will cause weakening of near-wellbore formation rock.

In the quest for production optimization from established resources, there is a continual interest in secondary and tertiary recovery methods. The success of these enhanced recovery methods, however, rely to a large extent on a sound understanding of fluid dynamics and migration paths in the reservoir. To this end, several approaches to reservoir characterization have been put to test with varying degrees of success. The unique ability of tracers to provide direct information on preferential fluid flow paths in the reservoir, and the sensitivity of partitioning tracers to fluid saturation distribution has highlighted the prospects of a detailed reservoir characterization through interwell tracer tests. In a broad sense, analysis of interwell tracer tests fall into two categories, analytical and inverse modeling. While most of the analytical methods are laden with limiting assumptions, the method of moments boasts rigorous formulation and accurate estimates of swept volume and average saturation of bypassed oil. The inverse modeling infers permeability and saturation distribution by matching the tracer response. An extremely effective approach to the inverse modeling methods computes sensitivities based on streamlines. The accurate modeling of tracer flow requires accounting for complex phenomena such as transverse dispersion. Also, it is sometimes desired to model pertinent tracer components through compositional simulation. This necessitates the inclusion of a physical dispersion tensor and, hence, the well established finite difference formulation. In this work, we have coupled the finite difference and streamline simulation techniques for the inversion-based reservoir characterization to take advantage of the robustness of the finite difference formulation and computational efficiency of streamline simulation. We have also extended the formalism of the inversion technique for fluid distribution estimation to scenarios with mobile oil saturations and have attempted integrating the analytical and inverse-modeling techniques to facilitate detailed reservoir characterization. We have demonstrated the feasibility of our approach on both synthetic and field cases.

The proposed work provides a new definition of the pressure derivative function [that is the β-derivative function, Δp βd(t)], which is defined as the derivative of the logarithm of pressure drop data with respect to the logarithm of time This formulation is based on the "power-law" concept. This is not a trivial definition, but rather a definition that provides a unique characterization of "power-law" flow regimes which are uniquely defined by the Δp βd(t) function [that is a constant Δp βd(t) behavior]. The Δp βd(t) function represents a new application of the traditional pressure derivative function, the "power-law" differentiation method (that is computing the dln(Δp)/dln(t) derivative) provides an accurate and consistent mechanism for computing the primary pressure derivative (that is the Cartesian derivative, dΔp/dt) as well as the "Bourdet" well testing derivative [that is the "semilog" derivative, Δpd(t)=dΔp/dln(t)]. The Cartesian and semilog derivatives can be extracted directly from the power-law derivative (and vice-versa) using the definition given above.

Accurately estimating hydrocarbon reserves is important, because it affects every phase of the oil and gas business. Unfortunately, reserves estimation is always uncertain, since perfect information is seldom available from the reservoir, and uncertainty can complicate the decision-making process. Many important decisions have to be made without knowing exactly what the ultimate outcome will be from a decision made today. Thus, quantifying the uncertainty is extremely important. Two methods for estimating original hydrocarbons in place (OHIP) are volumetric and material balance methods. The volumetric method is convenient to calculate OHIP during the early development period, while the material balance method can be used later, after performance data, such as pressure and production data, are available. In this work, I propose a methodology for using a Bayesian approach to quantify the uncertainty of original gas in place (G), aquifer productivity index (J), and the volume of the aquifer (Wi) as a result of combining volumetric and material balance analysis in a water-driven gas reservoir. The results show that we potentially have significant non-uniqueness (i.e., large uncertainty) when we consider only volumetric analyses or material balance analyses. By combining the results from both analyses, the non-uniqueness can be reduced, resulting in OGIP and aquifer parameter estimates with lower uncertainty. By understanding the uncertainty, we can expect better management decision making.

The Esdolomada Sandstone member 2 crops out in the Tremp-Graus Basin of north-central Spain and forms the uppermost part of the Eocene Roda Formation. The second Sandstone unit within the Esdolomada member (ESD2) consists of bioturbated and shell-rich, very-fine sandstones as well as stacked sets of fine- to coarse-grained cross-stratified sandstones. The overall upward trend in the member is commonly upward thickening and coarsening of beds into and through the cross-stratified interval, though at some few locations there is no obvious trend or even upward thinning of beds. The internal architecture of the member is one in which groups of beds lie between master surfaces that dip highly obliquely to the migration direction of the individual cross strata. The ESD2 is interpreted to be a shelf tidal sand bar within the overall transgressive Esdolomada Sandstone member. It is likely that these bars migrated in a coast parallel fashion, as suggested by the cross-bed orientations, but also accreted laterally away from the coast along the seaward-dipping master surfaces. LIDAR (light detection and ranging) data collection for the Esdolomada member was attempted along the Isábena River near the village of Roda de Isábena, with a total lateral coverage of approximately 3 kilometers. Detailed outcrop measurements were made in accessible areas along the same transect. Outcrop analogs are the best source of data to understand reservoir heterogeneities and to build reservoir analogs for fluid flow simulations. Sand-rich, offshore tidal sandbodies are usually surrounded by marine mudstones, and are recognized from their very orderly stacking of cross-stratified sets (more orderly than in fluvial settings) , their complex internal architecture of master surfaces dipping obliquely to the direction of migration of the contained cross strata and their significant sandstone/mudstone heterogeneities. Tidal bar systems such as the ESD2 are appealing hydrocarbon prospects for several reasons. Primarily, they are relatively coarse grained, have a high degree of lateral continuity, and are relatively clean sands. In places where sand beds are stacked, they create enough thickness to offer good vertical permeability; however, mud-draped cross-beds can create heterogeneities in this type of system that buffer fluid flow. Due to a fairly unsuccessful attempt to obtain LIDAR coverage of the ESD2, in order to build an analog reservoir model, surfaces were instead based on measured sections and outcrop photomosaics. Using Schlumberger’s Petrel software, facies logs were created from measured section data, and then interpolated to make a facies and porosity model. / text

A variety of oil recovery improvement techniques has been developed and applied to the productive life of an oil reservoir. Reservoir simulators have a definitely established role in helping to identify the opportunity and select the most suitable techniques to optimum improvement in reservoir productivity. This is significantly important for those reservoirs whose operating and development costs are relatively expensive, because numerical modeling helps simulate the increased oil productivity process and evaluates the performance without undertaking trials in field. Moreover, rapid development in modeling provides engineers diverse choices. Hence the need for complete and comprehensive case studies is increasing. This study will show the different characteristics of in-house (UTCOMP and GPAS) and commercial simulators and also can validate implementation and development of models in the future. The purpose of this thesis is to present a series of case studies with analytical solutions, in addition to a series of more complicated field cases studies with no exact solution, to verify and test the functionality and efficiency of various simulators. These case studies are performed with three reservoir simulators, including UTCOMP, GPAS, and CMG. UTCOMP and GPAS were both developed at the Center for Petroleum and Geosystem Engineering at The University of Texas at Austin and CMG is a commercial reservoir simulator developed by Computer Modelling Group Ltd. These simulators are first applied to twenty case studies with exact solutions. The simulation results are compared with exact solutions to examine the mathematical formulations and ensure the correctness of program coding. Then, ten more complicated field-scale case studies are performed. These case studies vary in difficulty and complexity, often featuring heterogeneity, larger number of components and wells, and very fine gridblocks. / text