Geological Sequestration as a Carbon Mitigation Option:
Capture of CO2 prior to atmospheric release, and subsequent injection in deep geological formations, is a promising option to reduce the rate of atmospheric CO2 increase. We are studying many aspects of this technology, known as Carbon Capture and Storage, or CCS. Our work focuses on the fate of the injected CO2 over both short and long time scales, the extent and magnitude of subsurface perturbations associated with the injection, and the potential leakages out of the injection formation of the sequestered CO2 as well as displaced resident fluids such as high-concentration brines. Our approach is mainly computational, although we are also involved in experimental projects and data analysis to help quantify fluid movement and leakage potential, with a general focus on leakage through existing oil and gas wells.
Multi-scale models for CO2 injection, migration, and leakage:
The CO2 problem is one that is often dominated by strong buoyant segregation of the less-dense CO2 and the more-dense brine. This segregation imposes a structure on the vertical profile of the fluid occupancy, which can be exploited in multi-scale modeling frameworks. We have developed a series of multi-scale models, taking advantage of vertical structure in the solution as well as localized features in the horizontal plane. These allow for very efficient simulation approaches, which we have used to study many aspects of the CO2 problem. The approach is explained in some detail in the recent book by Nordbotten and Celia (2012).
Leakage of CO2 and Brine along wells, faults and fractures:
Leakage of injected CO2, or possible resident fluids such as brine, is a major concern associated with injection of large amounts of CO2. Likely leakage pathways include old wells, and faults and fractures. In North America, millions of old oil and gas wells are co-located with the most suitable locations for CO2 injection. Therefore leakage along these small-scale features is important and needs to be included in models. Our multi-scale modeling framework is ideal to study these kinds of leakage problems, and new models based on these approaches have allowed us to analyze large-scale simulation domains (many thousands of square kilometers, with ten or more geological layers in the vertical) that include thousands of old wells, each of which can act as a leakage pathway. We continue to pursue this modeling work, with a current focus on faults and fractures.
Measurement of Methane Leakage from Abandoned and Orphaned Oil and Gas Wells:
Motivated in part by our earlier and ongoing work on potential leakage of injected CO2 along old wells, and in part by the need to quantify methane emissions into the atmosphere, we initiated a research program to measure methane leakage from old wells in northwestern Pennsylvania. More than 30 wells have been measured to date, using static flux chambers to estimate the methane fluxes. All wells are leaking some amount of methane, with about 15% of them being high emitters. Based on the average leak rate and an estimate of the total number of wells in Pennsylvania, we estimate the total methane flux to be about 10% of the total anthropogenic methane fluxes for the state of Pennsylvania. This source of methane is not included in any current emission inventories. Given the significant level of emissions, leakage from old wells should be included in future emissions inventories. The measurement program is continuing, with a focus on ways to identify the high emitters among the large number of wells.
CO2 Injection into Depleted Shale Gas Reservoirs:
Recent studies have suggested the possibility of significant geological CO2 sequestration in depleted shale gas formations. We are developing computational models to investigate the practicality of CO2 injection into shales. These models include two-component gas flow (methane and CO2) and appropriate representations of excess sorption for both components. Initial results show that while storage capacity is large, the rate at which the CO2 can be delivered declines rapidly in any given well. Several hundred injection wells would be needed to sequester the output from a single coal-fired power plant, with sequenced staging to maintain overall injectivity through time. In addition to the injection of CO2, we are also building models to simulate injection and subsequent imbibition and transport of fracking fluids into the shale system. The two-phase nature of the imbibition problem makes this simulation more computationally challenging.
Pore-scale models for two-phase flow in mixed-wet shale-gas systems:
Shale-gas systems are characterized by complex pore spaces with sizes in the nano-meter range, and by a water-wet rock matrix with hydrocarbon-wet kerogen inclusions. We are developing pore-scale network models, with regions of different wettability within the network, to study the behavior of the system. Our models include fluid compressibility, sorption, and flow modifications based on the very small pore sizes.