The introduction of materials such as CO₂ and radionuclides at non-geological rates initiates chemical, biological and mechanical interactions at conditions far from equilibrium. Thus, conditions in these subsurface systems likely differ significantly from those encountered in traditional hydrocarbon exploration and production settings that currently dominate our understanding of basinal fluid flow and storage processes. New tools and methods are necessary to characterize basin-scale features such as fractures and other discontinuities and relate their characteristics to processes at the molecular and pore scale.  

Figure. 1 Enhanced sequence characterization of a thin sand body mapping using improved seismic inversion. (a) Original seismic data and old interpretation. (b) Model based inversion impedance and new interpretation.

Non-equilibrium effects imply variation in process characteristics at different temporal scales. This necessitates new methods to analyze and reconcile core and field observations. A multi-tiered approach that includes outcrop and core-based structural and diagenetic investigations, seismic reservoir imaging, multiphysics modeling and reservoir monitoring to better characterize subsurface heterogeneity at the basin scale is being developed. A key focus is to understand and quantify the relationship of basin scale features to the observations at the smaller scales. .

How do distinct patterns of mineral alteration adjacent to fracture conduits affect sequestration capacity and hydraulic properties? 
 Fractures and faults are integral aspects of subsurface systems. Studying the long term impact of sequestered fluids on the mechanical and chemical properties of rocks becomes especially important in fractured systems.

What are the short and long-term effects of injecting fluids at non-geologic rates into fractured geologic systems and how can these effects be discerned using core and field observations?
The injection of fluids may induce short-term non-equilibrium effects that could modify the long-term physical and geochemical conditions of the confining systems especially in the vicinity of discontinuities inherent in the rock mass.

How can we enhance the accuracy and resolution of seismic imaging techniques so as to account for sub-seismic scale changes in rock properties?
A promising approach is to generate high-resolution probabilistic estimates of rock property variations and merge those maps with the low-resolution information inferred from seismic in order to image the progress of sequestration processes and quantify the associated uncertainty.

How can we statistically analyze patterns of fractures, diagenesis and other imprints of transport in geological systems and relate them to processes and characteristics at the smaller scales?
This would necessitate schemes for quantifying the statistical essence of features observed at the basin scale and subsequently deriving functional forms for these essential characteristics on the basis of features observed at the smaller scales. This is a major challenge, as many of the underlying processes are deterministic.

How do we model non-equilibrium processes that enhance or degrade the ability to sequester energy byproducts for long time periods? What are the critical physical and chemical processes and how are they coupled? How can uncertainty in predictive simulations of these coupled processes be characterized?
 Answering these questions requires developing models of fluid flow induced reactivation and/or sealing of faults and fracture zones. Application of new techniques for inverse modeling and quantification of uncertainty in the estimated parameters and development of new upscaling approaches are being done. This work has strong connections with Focus Area 4.

 

The research activities involve a combination of field and core based characterization of geological, chemical, and chemical processes. In chemically reactive environments such as CO₂ and other waste fluid reservoirs, chemical alteration processes and brittle deformation likely interact giving rise to complex feedback processes. We are analyzing the manifestations of these interactions using a combination of structural and diagenetic field mapping, petrographic and SEM imaging, mineralogical, elemental, fluid inclusion, and isotopic analyses, subcritical fracture testing, geomechanical modeling of fault-fracture interactions, and geochemical reaction and reaction path modeling.  

Field- and core-based studies are focused on surface-exposed sites of natural active CO₂ seepage (e.g. Little Grand Wash fault, Green River, Utah) and on CO₂ injection sites such as the Cranfield site. Outcrop and core based characterization are complimentary with outcrop based results providing insight into long-term (103 to 104 yrs) processes compared to core based (101 yrs) results. The outcrop component also provides structural and spatial insight not obtainable for subsurface datasets whereas core-based approach providing insight into fluid chemical boundary conditions and constraints on timing not readily available for outcrop analogs.

The characterization of reservoirs and processes at field scale will yield physics-based models that will form the basis for the numerical modeling effort described in Focus Area 4. The investigations of field scale effects of CO₂ injection and long-term storage will be integrated with, and build on, the grain-scale textural and mineralogical characterization of unaltered and altered reservoir rock outlined in Focus Area 2. 

• Reservoir-scale CO₂-brine/sandstone interaction in a natural CO₂ system: At the Crystal Geyser site in Utah, we find that CO₂-related alteration is strongest in the vicinity of hard-linked fault relays within the highly segmented fault system (Figure. 2). In addition, alteration is more extensive at topographically lower fault relays suggesting that CO₂ seepage along the Little Grand Wash fault reflects mixing of CO₂-ladden saline basinal fluids with meteoric water. This view is consistent with preliminary stable isotope analyses of carbonate sandstone cement, showing trends of heavier oxygen and carbon isotopes in the vicinity of fossil CO₂ seeps. 

 

Figure 2. Upper: Graduate student Alex Urquhart sampling a calcite vein in CO₂-altered sandstone along the Little Grand Wash fault, Utah. Lower:  Distribution of travertine and CO₂-altered sandstone within a highly segmented portion of the Little Grand Wash fault, Utah.   

Effective permeability of heterogeneous fractured and faulted materials: A cellular automata (CA) model for the pressure distribution within a planar fault zone having heterogeneous material properties (compressibility and porosity) has been developed. The effects of spatial correlation in material properties on the resultant behavior are being studied (see Figure. 3). Additional work will examine redistribution of pressure due to episodic introduction of fluids into the fault zone.

Inversion of reservoir parameters at the Cranfield Field injection site: The project focuses on answering the key questions: 1) How will a fault in a reservoir undergoing CO₂ injection perform? Are monitoring techniques sensitive enough to detect leakage. We have developed a new paradigm of geological model selection based on observed injection history data that will help answer questions such as locations of sub-seismic scale faults and their effect on CO₂ plume movement. 

                                              

Figure. 3. Effect of spatial correlation of rock compressibility on pressure redistribution within a fault zone over time from uniform stress. Images show the pressure distribution at three times for an uncorrelated compressibility field (left column) and a compressibility field with isotropic correlation length defined by a Gaussian kernel with a full-width at half maximum (FWHM) of 18.4 pixels.

Understanding the mechanisms for periodic eruption of carbon-dioxide driven, cold-water geysers: Periodic eruptions of CO₂-laden water have been observed at abandoned well sites in the Crystal Geysers area in Utah. Our focus has been on setting up a simple tank-type model for the process that incorporates the phase behavior of CO₂-water binary mixtures under quasi-isothermal conditions and a recharge mechanism of CO₂ and water into the tank.

3D characterization of Dickman CO₂ Sequestration site: The central focus of our study is to characterize a subsurface reservoir in three dimensions that is a likely candidate for carbon storage. 3D seismic and well log data are being used to estimate not only the acoustic impedance but also the shear impedance and density. A 3D porosity model (Fig. 4) was developed which will be further improved by using full waveform inversion.

Figure 4: 3D display of porosity estimate along the target zone derived by using a multi-attribute regression analysis of pre-stack inversion results. Note the high porosity zone and its extent around well Elmore 3.

Scale-Up of Transport Processes in Heterogeneous Reservoirs: A new technique to quantify the scaling characteristics of transport processes based on the volume averaging approach has been developed. We studied the scaling characteristics of effective mass transfer coefficient for a tracer injection process in single-phase flow corresponding to different reservoir heterogeneity correlation lengths as well as different transport mechanisms (e.g., convection, dispersion, and diffusion). The scale-up of active tracer partitioning between two movable phases was studied and results (Figure. 5) indicate that differences in scaling characteristics of Keff become apparent when the degree of heterogeneity increases (e.g., longer correlation lengths), especially in convection-dispersion dominant processes.

                        

Figure 5: Scaling characteristics of breakthrough time (BT) and tracer recovery at base case conditions for the case of tracer transport from water phase to oil phase.

Figure 6. A heterogeneous permeability field in a 3D sandbox (21 x 9 x 8.5 cm3) was constructed with five different size of sands and purged with a solution of tracer uniformly. Measured tracer signal using MRI at a voxel scale (~0.25x0.25x0.25 cm3) at a regular interval (130 s) time over 4 hrs was converted to concentrations. This experiment serves as a test bed for inverse parameter estimation in heterogeneous fields.

 

Events                                                                                          top

 

Publications, Abstracts, and Posters                                        top

 

News                                                                                            top