This study measured the liquid fallback during simulated blowout conditions. The purpose of the study was to establish a basis for developing a procedure for controlling blowouts that relies on the accumulation of liquid kill fluid injected while the well continues to flow. The results from experiments performed with air, water, 10.5 ppg and 12.0 ppg mud in an experimental 48 ft flow loop at 0°, 20°, 40°, 60° and 75° deviation angles from the vertical, as well as results from full-scale experiments performed with natural gas and water based drilling fluid in a vertical 2787-foot deep research well, are presented. The results show that the critical velocity that prevents control fluid accumulation can be predicted by Turner's model of terminal velocity based on the liquid droplet theory by also considering the flow regime of the continuous phase when evaluating the drag coefficient, and the angle of deviation from the vertical. Similarly, the amount of liquid that flows countercurrent into and accumulates in the well can be predicted based on the concept of zero net liquid flow (ZNLF) holdup. Finally all these concepts are integrated in the dynamic kill procedure, which is based on system performance analysis to better predict the feasibility of an off-bottom dynamic kill.

Approximately 2.5 billion dollars is spent annually to solve the problem of produced water in oil and gas wells. Downhole Water Sink (DWS) technology is one industry solution to control water coning in oil wells. DWS technology involves the segregated production of oil and water through separate completions with zonal isolation packer. However, several problems have been experienced in the application of the technology in watered-out oil wells. This study identified two factors that could aid in a better modeling of the technology in old vertical wells inclusion of capillary transition pressures and relative permeability hysteresis. It also identified a pressure enhanced capillary transition zone enlargement around the wellbore as responsible for the concurrent production of contaminated fluid from both completions. Another widely recommended industry solution to the problem of produced water is horizontal well technology. However, field reports indicates that water breakthrough into horizontal wells could be quite dramatic and tend to erode the merit of high deliverability. This study analyzed the problem of water cresting in horizontal wells and developed a generalized compound friction pressure loss relation for horizontal wells and pipes. The new relation includes factors such as perforations, oil-water emulsions, and radial influx of fluid into the wellbore as well as phase inversion. It also shows the results of the application of this relation in the modeling of water cresting in horizontal wells subject to bottom water drive. These results reveal an asymmetrical distribution of water influx skewed toward the heel in line with field observations. Finally, the study presents two innovative dual-completion concepts for controlling water cresting in horizontal wells adapting the principles of the Downhole Water Sink technology. The results of the initial studies shows that oil recovery could be improved by as much as 7 percent over conventional horizontal wells.

Dimensionless space provides a tool for analyzing the behavior of complex systems described by mathematical relationships. The limited application of dimensionless variables in numerical reservoir simulation and experimental design motivated the development of a complete set of dimensionless scaling groups. Inspectional analysis yielded 8 dimensionless groups completely describing the flow system. Further analysis of fluid interaction reduced the number of dimensionless groups to 7. The newly developed dimensionless equations and groups were used for analytical and numerical reservoir characterization, quantifying the behavior of differential and difference equations employed in fluid flow in three-dimensional porous media. The behavior of the dimensionless scaling is demonstrated for breakthrough time in an immiscible displacement in three dimensions. Numerical simulations were designed in dimensionless space and converted to dimensional space using several approaches. The resulting estimates of stability limits, numerical dispersion, and regime boundaries were in excellent agreement. The application of the dimensionless groups to upscaling was investigated using designed reservoir simulations to estimate dimensionless regions corresponding to different flow regimes. Analytical development, simulation runs and literature data were in good agreement. This application demonstrates the potential benefits of the proposed dimensionless groups for upscaling, sensitivity analysis, stability analysis, and reservoir characterization.

Safety culture relating to offshore operations has shifted since the Deepwater Horizon blowout and resulting oil spill. This incident has prompted the research of high volume spills during all stages of hydrocarbon exploration and production. This study particularly covers the interactions of wells and offshore networks as they pertain to situations where a release of reservoir fluids to the environment is occurring. Primary concerns of this investigation are stream confluences, leak modeling, and fluid behavior; the first two will be handled with various numerical software packages (OLGA®, CFD, and nodal analyses) while the later will require more rigorous treatment and a combination of these tools with dedicated phase behavior software (such as PVTsim®). This research will combine with risk analysis work being done by others to identify high-priority system failure scenarios. The focus in modeling high-volume leaks thus far has been placed upon reservoir properties, geology and modeling the most uncertain things when this research shows that the most influential variables for particular reservoirs lie within the flow path. When operating offshore, wells connect to subsea manifolds or other junctions to form unforeseen mixtures of crude oils; these combined fluids dictate the outcome of potentially devastating releases offshore. Flow rates through chokes have been modeled using only a few parameters, namely the pressure, choke size and the gas-liquid ratio (GLR). The leak considered herein will choke flow and create a back pressure, which will control how fluids move from the reservoir to wellhead. A properly tuned equation of state can predict the GLR fairly well, but falls short when attempting to combine the GLR of two or more fluids. A correlation is proposed to allow for more accurate leak models when only simple fluid properties are known, such as the heptanes-plus fraction.

Transporting drilled cuttings from the bottomhole to the surface becomes more difficult and problematic in highly deviated wells than in vertical wells. Cuttings tend to settle down on the low side of the annulus typically in the form of a bed which can cause further problems. The height of this bed depends on many parameters such as annular domain geometry, drilling fluid density and rheology, annular flow rate, drill pipe rotation speed, cuttings size, shape, and their density. Prediction of the stationary cuttings bed height with respect to these aforementioned parameters is thus necessary to optimize the range of the controllable parameters for a desired level of wellbore cleaning. A computational setup that represents the spatial geometry of the cuttings transport domain, and utilizes discrete phase model coupled with numerical solution of the Navier-Stokes equations augmented by a turbulence closure model SST version of k-ω is used for predicting the bed height of the stationary cuttings bed and moving cuttings velocities. Discrete phase model is a mathematical tool to navigate large number of particles in a flow field by calculating the particle paths in a Lagrangian frame by the time integration of force balance on each individual particle. Turbulence effects on the particle motion are also incorporated through a random walk model. The drag force on non-spherical particles is incorporated using a sphericity based correlation. Roughness of stationary bed surface is also incorporated through the modified law-of-the wall model. A snapshot technique is applied here in these simulations by computing flow solutions in the geometric domains with pre-defined stationary bed heights. The statistics of particle - wall collisions are analyzed over these geometrically pre-defined stationary bed surfaces to predict which domain would represent the equilibrium cuttings bed height. A systematic validation study is presented by comparing the simulation results against published experimental datasets for velocity profile estimation of non-Newtonian fluids flowing in turbulent regime, stationary bed heights, and moving bed velocities. Further, a parametric study is presented for the effects of wellbore inclination, fluid density and rheology, particle size and sphericity, inner pipe rotation and the inner pipe rotation speed.

In todays exploration and production environment it is required that engineers must collect and use vast amounts of data for flow model construction and calibra- tion, as well as reservoir estimation and optimization. With modern technology, the data volume can be overwhelming. It is necessary that data monitoring and calibration are highly efficient. It is also essential that physical and mathematical models can be tested in repeatable, inexpensive experiments. The experiments described in this thesis will develop and perform verification of algorithms and can generate prior geomodels, collect and process seismic refraction data, collect and process production data, and calibrate these models. The experimental components discussed here are collectively referred to as The Sand Tank Experiment. Contained in the LSU WaveCIS tank is a wedge shaped sand pack that can be saturated with water. Water can then be produced from this model reservoir while it is monitored by pressure/temperature sensors. A 20 kHz seismic source and 8 accelerometers are used to collect seismic first arrival data during this production period. This data can then be used to image varying water saturations throughout the reservoir. Those water saturations modify the compressional, p-wave, seismic velocities as described in the Biot and Gassman relationships. Picking first arrival times for each run of the experiment can further enhance the use of the seismic data. These first arrival times can then be compared to calculated first arrival times from simulation data and the residuals can be used to measure the accuracy.

Realtime reservoir management is developed to manage a shrinking labor force and rising demand on energy supply. This dissertation seeks good strategies for realtime reservoir management. First, two simulatorindependent optimization algorithms are investigated: ensemblebased optimization (EnOpt) and bound optimization by quadratic approximation (BOBYQA). Multiscale regularization is applied to both to find appropriate frequencies for well control adjustment. Second, two gathered EnKF methods are proposed to save computational cost and reduce sampling error: gathered EnKF with a fixed gather size and adaptively gathered EnKF. Finally, oil price uncertainty is forecasted and quantified with three price forecasting models: conventional forecasting, bootstrap forecasting and sequential Gaussian simulation forecasting. The relative effect of oil price and its volatility on the optimization strategies are investigated. A number of key findings of this dissertation are: (a) if multiscale regularization is not used, EnOpt converges to a higher net present value (NPV) than BOBYQAeven though BOBYQA uses second order Hessian information whereas EnOpt uses first order gradients. BOBYQA performs comparably only if multiscale regularization is used. Multiscale regularization results in a higher optimized NPV with simpler well control strategies and converges in fewer iterations; (b) gathering observations not only reduces the sampling errors but also saves significant amount of computational cost. In addition, adaptively gathered EnKF is superior to gathered EnKF with a fixed gather size when the prior ensemble mean is not near the truth; (c) it is shown that a good oil price forecasting model can improve NPV by more than four percent, and (d) instability in oil prices also causes fluctuation in optimized well controls.

<p>Well cementing is one of the key processes performed during drilling and completion of wells. The main objective of primary cementing is to provide zonal isolation. Failure of cement to provide zonal isolation can lead to contamination of fresh water aquifers, sustained casing pressure, or blowout. For effective cementing, the cement slurry should completely displace the drilling fluid. In practice, this is hard to achieve and some mud is left in the wellbore where it contaminates the cement after cement placement.</p> <p>This study investigates the effect of mud contamination of cement on the integrity of cement-formation interface. Flow-through experiments were conducted over 30-day periods using cement-sandstone composite cores and brine (salinity ~20,000 ppm) at 2100 psi (14.48 MPa) confining pressure, temperature of 72°F (22.22°C) and flow rate of 1 ml/min. Each cylindrical composite core was composed of half-cylinder Berea sandstone and half-cylinder Portland cement paste, with dimensions 12 in (30.48 cm) in length and 1 in (2.54 cm) in diameter. One composite core had no contaminated layer and each of the two other cores had a ~1.27 mm (0.05 in) thick layer of contaminated cement with 5% or 10% mud contamination by volume.</p> <p>Image based techniques used to characterize the composite cores revealed the presence of large non-circular pores (with as much as ~750 µm length and ~150 µm thickness) in the mud contaminated cement at the end of the core-flood. The large pores were higher in number in the 10% than in the 5% mud contaminated cement. The pores did not appear to be interconnected at the end of 30-day core-flood although leaching of the cement surrounding the large pores was observed. The porosity of the 10% mud contaminated cement was found to have increased from 0.65% to 12.53%.</p> <p>Based on the observations, we can conclude that large pores are created in cement due to the presence of mud contamination and the pores increase in number as level of mud contamination increases. Leaching of the cement surrounding the large pores may lead to interconnectivity of the large pores in the long run and result in loss of zonal isolation.</p>

Geothermal energy has gained a lot of attention recently due to several favorable aspects such as ubiquitously distributed, renewable, low emission resources while leveraging the advances in the associated technologies such as directional drilling and low enthalpy power generation plant. However, there are still many challenges such as the high initial capital cost of drilling and surface facilities, environmental risk of seismicity due to the induced disequilibrium in the formation, and sustainability of project over designed operational life. Traditional downhole heat exchangers (DHE) could potentially reduce the capital cost and the risk of seismicity, but they are unable to maintain a sustainable geothermal energy production over the operational life due to the rapid cooling down of formation in the vicinity of the wellbore. In this study, a novel DHE design is introduced to enhance the energy production rate as well as sustainability for mainly two types of geothermal reservoirs: saturated geothermal reservoirs and enhanced/engineered geothermal systems (EGS). Modeling of DHE is based on the concept of thermal resistance. A geothermal reservoir simulator is built reusing components of an existing blackoil simulator by adding thermal energy transport equations and fracture representation (discrete fracture network). Several verification and validation tests are carried out. Parametric studies are presented for various configurations of DHE and thermodynamic analysis is carried out for the binary power plant cycle. In addition, the geothermal reservoirs Camerina A and Raton Basin are presented as case studies for saturated geothermal reservoir and EGS, respectively. In saturated geothermal reservoirs, the performance of DHE is improved significantly by exploiting forced convection. For EGS, the overall heat extraction rate is also enhanced by adding DHE.

More than sixty percent (60%) of conventional hydrocarbon reservoirs which are potential CO2 repositories are sealed by tight shale caprock. The geochemical reactivity of shale caprock during CO2 diffusive transport needs to be included in the reservoir characterization of potential CO2 sequestration sites as slow reactive transport processes can either strengthen or degrade seal integrity over the long term. Several simulation results had predicted that influx-induced mineral dissolution/precipitation reactions within shale caprocks can continuously reduce micro-fracture networks, while pressure and effective-stress transformation first rapidly increase then progressively constrict them. This experimental work applied specific analytical techniques in investigating changes in surface/near-surface properties of crushed shale rocks after exposure (by flooding) to CO2-brine for a time frame ranging between 30 days to 92 days at elevated pressure and fractional flow rate. Initial capillary entry parameters for the shale were estimated from digitally acquired pressure data evolution. Flooding of the shale samples with CO2-brine was followed by geochemical characterization of the effluent fluid and bulk shale rock through ICP-OES, XRD, EDS and pH measurements. Nano-scale measurement of changes in internal specific surface area, pore volume and linear/cumulative pore size distribution (using the BET Technique) showed that changes in the shale caprock due to geochemical interaction with aqueous CO2 can affect petrophysical properties. The intrinsically low permeability in shale may be altered by changes in surface properties as the effective permeability of any porous medium is largely a function of its global pore geometry. Diffusive transport of CO2 as well as carbon accounting could be significantly affected over the long term. The estimation of dimensionless quantities such as Peclet (Pe) and Peclet-Damkohler (PeDa) Numbers that are associated with geochemical reactivity of rocks and acidic fluid transport through porous media gave insight into the impact of diffusion and reaction rate on shale caprock in CO2 sequestration.