Ye Zhang, PhD , University of Wyoming
Injection of carbon dioxide (CO2) into deep saline aquifers is considered a promising option to mitigate global climate change. To assess the appropriateness of a storage site, reservoir simulation is commonly performed. As part of this analysis, a geologic site model must be built. Since CO2 sequestration into saline aquifers is a “cost center”, research is needed to develop a cost-effective strategy in data collection to support the building of the site models. This research aims to identify important uncertainty factors that impact CO2flow prediction in a deep saline aquifer. Since high-resolution mapping of aquifer intrinsic permeability (k) heterogeneity is prohibitively expensive, the importance of resolving detailed heterogeneity is evaluated. In particular, we aim to identify an optimal resolution in an aquifer model with which sufficiently accurate predictions can be made.
2. Summary of Research Activity
We started with a three-dimensional synthetic aquifer where we developed multiple conceptual models at different heterogeneity resolutions. We found that models with coarsened heterogeneity resolutions can indeed capture certain aspects of CO2flow dynamics, although the optimal resolution depends on the prediction goal and the strength of heterogeneity. This work has produced one published article and one article in press.
In Spring, 2011, we proceeded to conduct the second phase of our study, that is, to test the insights gained from the synthetic system by modeling a natural deep saline aquifer. This aquifer is the basal Mt. Simon Sandstone in the Illinois Basin which has been selected as a prospective CCS site due to its successful history as a natural gas storage formation and accessibility to large CO2 point sources. A large CO2 capacity is also projected for Mt. Simon which consists of quartz sandstone interbedded with thin siltstone and shale. From limited wireline logs and outcrop exposures, it is known to be heterogeneous. At Mattoon, the proposed injection site of the FutureGen project, Mt. Simon lies at depth ~ 2km within a tectonically quiescent region. A single-injection-well CO2 demonstration project was scheduled in 2011. With cores extracted from Mt. Simon and its caprock, in-situ experiments will be conducted to evaluate brine-CO2-rock interactions. CO2-brine relative permeabilities and capillary pressure will also be measured. Before injection, baseline surveys will record the hydrogeological, geochemical and geophysical conditions within the reservoir and its caprocks. During injection, dynamic information will be gained, e.g., pressure monitoring, fluid samplings, and seismic surveys. After injection, monitoring techniques will be employed to verify CO2 storage and provide early warnings should leakage occur. The test results, along with data on site geology/geophysics, will be in public domain based on which we will conduct an independent model evaluation, our results compared to those of other modeling teams. Specifically, we will construct a geologic site model at multiple resolutions (e.g., formation-homogeneous, facies-zonation, fully heterogeneous), conduct reservoir simulations based on the computationally efficient Design of Experiment approach, and assess the uncertainty in CO2 predictions for the different conceptual models. We will again aim to identify an optimal geologic site model with which predictions will be made, which will be compared to the results of the field test. This comparison will serve to validate the optimal site model(s).
In the past 6 month since we initiated the second phase of our study, our research has focused on 3 aspects:
(1) comprehensive literature review of past works and relevant modeling reports:
Our literature review has compiled past modeling projections on CO2plume migration, pressure evolution, and their impact on the basin’s deep and shallow groundwater systems (e.g., extent of saline water migration due to pressure buildup at the injection site). We plan to compare our model outcomes with those of the previous studies, although most of them did not build detailed reservoir models.
(2) collection of public-domain characterization and monitoring data of the Mt. Simon Sandstone:
Our effort has yielded both regional and local data: (a) regional data: 5 maps, 4 cross sections, 50 well logs, and a large set of core porosity and permeability measurements; (b) local data: a set of interpreted seismic horizons near the Mattoon injection site.
(3) building a geologic structural model of the Mt. Simon Sandstone:
Due to the great burial depth of Mt. Simon, especially in and around the basin center, to date, only 68 wells have been drilled into this formation. We collected wireline logs from 27 of them, which were released to us from the Illinois State Geological Survey (ISGS). (We had also contacted the FutureGen Consortium, but their data had not been deemed suitable for release at the time of our communication. We plan to contact them again in the near future.) Most of these wells are not located near the Mattoon injection site which lies in the middle of the basin. Instead, they are distributed throughout the Illinois Basin. For the 27 wells, we have digitized ~50 wireline logs covering the Mt. Simon interval which will yield information on formation structure, facies, petrophysical properties (porosity, permeability), and fluid type. For example, at one well, we have neutron porosity, resistivity, salinity, saturation, and lithology (see TOC). However, many wells have much more limited log suites: only 6 among the 27 wells have sonic or neutron density logs from which formation porosity can be inferred. All wireline logs were also inspected for accuracy to identify potential inconsistencies in well location or depth. Due to sparseness in well location, interpolation of formation structure at the regional scale was problematic using well logs alone. Thus, we constructed a structural model of the Mt. Simon Sandstone, combining well log formation picks and published contour maps of Mt. Simon’s top and bottom surfaces. From published sources, we have also identified 11 faults that cut through Mt. Simon, extending along NW-SE with high dip angles approaching vertical. Some of these faults, due to their proximity to the injection site, could have important effects on CO2flow and pressure buildup within Mt. Simon and its caprocks, if fault permeability is significant. Thus, these faults have been built into the structure model (see TOC).