Projects

Grain-scale to continuum modeling and experiments of fluid flow and solid deformation in unconsolidated, heavy oil sands: Applications to injection well fracturing and near-wellbore deformation

Principal Investigators: Steve Bryant, Jon Olson and Mukul Sharma

Field and laboratory data clearly show that fracturing in poorly consolidated rocks is not adequately represented by traditional elastic models. It is the objective of this proposal to investigate the mechanisms of fracture propagation in unconsolidated sands. This project will focus on the coupled flow and deformation problem for unconsolidated sands from three different but complimentary perspectives: i) grain-scale modeling, ii) continuum application modeling, and iii) laboratory experiments.

Grain-scale modeling will focus on understanding fundamental deformation and flow behavior, building on previous research with the commercial codes PFC2D and PFC3D which established a grain-level basis for the fracturing and dilation behavior of sands which have undergone different degrees of consolidation. We will integrate with these mechanics codes a powerful method for a priori calculation of fluid pressures at the grain scale which will significantly increase the physical validity of the predicted mechanical behavior. The model will account for the effects of temperature and temperature gradients explicitly within the fluid and solid constituents. We will use the code to evaluate the nature and the extent of reservoir fracturing and wellbore failure induced during typical heavy oil operations (cold production, steam injection and cycling, solvent injection). The behavior observed in ensembles of 104 to 105 grains will used to inform continuum-scale relationship that would be suitable for simulations of a single well or a reservoir.

The proposed continuum modeling will initially focus on fracturing behavior around injection wells, departing from traditional elastic techniques to account for the substantial non-elastic (plastic) deformation expected in weak sand formations. The initial approach will focus on the propagation of pore pressure and the resultant effective stress changes that alter rock porosity and permeability. It is postulated that a region of enhanced porosity defines a "fracture" in unconsolidated sands, and the nature of this deformation zone as described from the fundamental grain-scale work will be incorporated into the continuum model for large scale application.

Experiments will be conducted to measure the permeability of unconsolidated sands under a range of anisotropic loading conditions (up to and including incipient failure). By carefully controlling the properties of the unconsolidated sands and the stress state, the mechanisms by which zones of failure propagate in sands will be studied. The effect of the presence of a heavy-oil phase on the failure and flow properties of the sand will also be studied. In the first 12 to 18 months of work we will focus on the following activities:

• Coupling our research code that computes fluid pressures at the grain-scale with the commercial code PFC3D that computes grain movement and deformation
• Use the coupled grain scale simulator to predict the relationship between permeability and stress/strain of an unconsolidated granular material
• Validate grain-scale predictions against a series of experiments on poorly consolidated porous media to be run in our triaxial testing apparatus
• Develop a constitutive relationship for permeability/stress/strain suitable for inclusion in a continuum model
• Implement the constitutive relationship into a coupled geomechanics/fluid flow reservoir simulator already being developed at UT

Development of a general framework for scaling up recovery mechanisms and processes

Principal Investigators: Sanjay Srinivasan, Larry Lake, Steven Bryant

Evaluation of oil recovery processes is generally performed in laboratories using sand packs and/or rock cores. Extrapolation of laboratory results to pilot scale and eventually field scale process performance necessitates an adjustment or scale-up of the laboratory measured values. Several researchers have reported a decline in recovery efficiencies when a process established to be viable at the laboratory scale is applied in the field. Poorly understood scale up characteristics of reservoir attributes that are heterogeneously distributed in space and self- or autocorrelated and the complexity of processes that intrinsically occur at the pore level and are the culmination of several physical and chemical mechanisms contribute to the difficulty in accurately scaling up laboratory measured values to field scale quantities.

The proposed research will initially investigate the scale up characteristics of static reservoir attributes such as effective permeability. Suitable modifications to the volume-support relationships that are routinely used to represent linearly averaged processes will be attempted taking into account the non-linearity in the relationship between permeability and flow response. We propose to study the phenomena of pore-level mixing with the objective of understanding the scale up of dispersion. The influence of pore network connectivity on mixing over different volume supports will be studied so as to link the scale up characteristics of dispersion to the heterogeneity characteristics of the reservoir.

Current techniques for analyzing the scaling characteristics of displacement processes in porous media include homogenization (or volume averages) and ensemble averaging (or method of moments). Both these techniques are applicable to cases where the attribute of interest varies gradually at the fine scale and when the length scale of the averaging volume is significantly larger than the correlation range of the attribute at the fine scale. In the case of most oil recovery processes, mechanisms at the local scale (reaction) work together with mechanisms that are non-local (dispersion). The criteria for application of either the homogenization principles or ensemble averaging may not strictly hold. Our initial research focus will be to develop functional groups that quantify the local and non-local effects associated with typical recovery processes e.g. steam flooding. The scaling characteristics of these functional groups on a stand-alone basis will be evaluated first. Subsequently the influence of coupling on the scaling characteristics of the functional groups will be assessed.

Thermally enhanced chemical recovery of heavy oil

Principal Investigators: Gary Pope, Quoc Nguyen, Chun Huh

The general concept in this proposed research is to investigate the combined use of heat and chemicals to recover heavy oil. The chemicals might include surfactants, alkali such as sodium carbonate and solvents. Some of our earlier work with a 1000 cp oil resulted in an effective recovery process using a tailor made surfactant, alcohol co-solvent and a slightly elevated temperature of about 120 F. The increase in temperature both decreased the oil viscosity and promoted the formation of a microemulsion that would otherwise not have been feasible. The viscosity of the microemulsion was only about 3 cp. We also know from previous research that adding sodium carbonate to the surfactant solution can reduce the surfactant adsorption to nearly zero. It may also have other beneficial effects including reactions with the heavy oil and the enhancement of mass transfer. For oils with a viscosity greater than 1000 cp, it may be necessary to increase the temperature to 160 F or perhaps even higher. We propose conducting laboratory experiments to explore various ideas along these lines. We will use novel surfactants based upon our long experience with tailoring surfactants to various oils. We will also explore the mechanisms involved and develop conceptual models of potential processes under different conditions.

Investigation of methods to improve SAGD process

Principal Investigators: Quoc Nguyen, Larry Lake, Chun Huh, Sanjay Srinivasan

Steam flooding, SAGD, and cyclic steam stimulation have been employed worldwide as successful thermal methods for heavy oil recovery. The efficiency of these processes has been continuously improved by taking advantage of non-condensable gases for controlling heat loss, and enhancing the reduction of oil viscosity with addition of condensable solvents. Different injection patterns and operating strategies have been introduced to strengthen the synergistic effects of these combined processes. Nevertheless, the current state of the art still suffers from the influence of inherent geological heterogeneity and poor solvent mass transfer, frequently leading to high operational cost and poor sweep efficiency. To remedy these shortcomings, we propose a new process in which the distribution of steam-solvent is optimized using a non-conventional diverting technique. With strategic placement of diverting agents (e.g., microgels whose swelling can be controlled by temperature, pH or salinity changes; steam foam which can also carry solid particles for temporary filter-cake buildup; and viscoelastic surfactants whose viscosity could significantly change in the presence of hydrocarbon), we aim to optimize the steam chamber development in a designed manner. Of importance, the heat loss to overburden could be mitigated by placing non-conventional foam as an effective thermal "insulator" between the overburden and the steam chamber. Additionally, a novel solvent system is developed to suppress the formation of condensed-water front, thus enhancing solvent mass transfer. Viscous fingering of steam/solvent into bitumen is also deliberately utilized as an enhanced mass and heat transfer mechanism. To this end, we propose a parallel program of experiments and modeling. We will conduct experiments on the coupled thermal and solvent-assisted process of heavy hydrocarbon mobilization using CAT scanning for visualizing the in-situ fluid distribution and flow profiles. In parallel to the experiments, a mechanistic model will be developed to interpret the experiments, and as a first step towards scale up of the new processes.