2013 Carbon Management Technology Conference

Constructing a Reservoir Model to Simulate Commercial Scale CO2 Injection and Optimization of Storage Capacity in the Southeastern United States

Authors

Jonsson, H. - Presenter, Advanced Resources International
Cyphers, S., Advanced Resources International, Inc.
Petrusak, R., Advanced Resources International

This project explores the optimization of commercial CO2 injection and storage within the vertical geologic continuum from depth to surface in the Gulf Coast region of the southeastern United States. The Gulf Coast region is underlain by a massive “wedge” of predominately clastic deposits that spans the region and thickens basin ward into Gulf of Mexico geologic basin.  Consequently, a thick sequence of potential storage reservoirs and sealing formations will be present at any potential CO2 storage site selected because of favorable geologic structure.

The Commercial Scale CO2 Injection project leverages the site-specific geologic data available for the SECARB Anthropogenic Test Site at Citronelle Field in Mobile County, Alabama to fully quantify and optimize the storage capacity and injectivity of each reservoir unit at the site.  This paper evaluates a potential sub-surface storage volume that occurs from a depth of approximately 12,000 vertical feet to surface, and over a geographic area of more than 30 square miles.  This paper reports on two components of the project: 1) reservoir characterization of multiple potential storage and confining layers and construction of geologic models for each storage reservoir unit and, 2) characterization of reservoir flow unit heterogeneity and the appropriate up-scaling of reservoir flow properties to adequately represent this heterogeneity in the reservoir simulation. 

Modern geophysical logs and core data obtained from the new injection and observation wells drilled at the Anthropogenic Test Site were used to extrapolate porosity and accurate reservoir thicknesses from the 1960’s vintage induction electric logs at Citronelle Field.  Neural network methods were applied to extrapolate gamma ray response, bulk density and porosity from the vintage spontaneous potential and resistivity logs.  Reservoir permeability was extrapolated from porosity based on new core data obtained at the Anthropogenic Test site and the SECARB Phase II Saline Formation Test site plus other data available from the Geological Survey of Alabama.  Geostatistical methods were used to extrapolate areal reservoir heterogeneity from the vertical reservoir heterogeneity observed in well logs to obtain a fine scale three-dimensional rendering of reservoir storage volume and flow properties.  Lastly, the fine scale geologic characterization was aggregated or ‘up-scaled’ to specify representative flow units and confining layers for each formation.  The flow units may correspond to a single sandstone or to one of multiple permeable zones within sandstones.  More than three hundred representative flow units, or model layers, ranging in thickness from 10 to 100 ft. were prepared for three-dimensional visualization and subsequent input to the reservoir simulation. 

Subsequent papers will report on the simulation of various CO2 injection scenarios, fate and transport modeling of injected CO2, and three-dimensional visualization of the model results.