Modeling and Statistical Assessment of Fluid Migration Response in the Above Zone Monitoring Interval of a Geologic Carbon Storage Site
Carbon dioxide (CO2) capture and storage (CCS) into geological formations is regarded as an important strategy for achieving a significant reduction in anthropogenic CO2 emissions. An increasing emphasis on the industrial-scale implementation of CO2 storage in geological formations has led to the development of whole-system models to evaluate the performance of candidate geologic storage sites and the environmental risk associated with them. The United States Department of Energy (DOE), through its National Risk Assessment Partnership (NRAP) Program, is conducting research to develop science-based methods to quantify the likelihood of risks associated with the longterm geologic storage of CO2. A key component of this research is the development of an Integrated Assessment Model (IAM), which simulates the geologic storage system by integrating the primary sub-system components of the system for the purpose of elucidating the relationship between CO2 injection into the subsurface and the short- and long-term containment of the stored CO2. The sub-system components of that engineered geologic storage system include the storage reservoir, overlying aquitards (primary caprock and secondary seals) and aquifers (including the above zone monitoring interval, or AZMI, which directly overlays the primary seal), and potential leakage pathways including wells, fractures, and faults, to name a few. The subsystems are modeled using simplified reduced order models (ROMs), which are then coupled together to characterize the entire storage system. The construction of the IAM permits the quantification and assessment of the storage risks in a more computationally efficient manner as compared to the use of full physics-based models. This Ph.D. research was initiated with a review of the IAM development efforts of NRAP and an assessment of the current status of the ROMs for each of the sub-system components of the storage system. In addition, gaps in the development in the IAM structure were identified. The research also investigated the potential migration of subsurface fluids (i.e., CO2 and formation brine) in geologic storage sites. Leakage of CO2 and brine through the primary seal to the overlying porous and permeable formations, such as the AZMI, may occur due to the intrinsic permeability of the seal and/or the presence of natural fractures or induced perforations or fractures. Pressure modeling of the AZMI provides a useful source of information regarding seal performance and the subsurface pressure response to CO2 leakage from the reservoir. As part of this research, a ROM for the AZMI, a missing component of the current IAM of NRAP, was developed. This AZMI ROM simulates fluid flow above the primary seal and predicts spatial changes in pressure over time due to this fluid migration from the reservoir. The performance of the AZMI ROM was verified using full physicsbased models of the zone. A data-driven approach of arbitrary Polynomial Chaos (aPC) Expansion was then used to quantify the uncertainty in the pressure predictions in the AZMI based on the inherent variability of the different geologic parameters such as the porosity, permeability, and thickness of the AZMI, and the caprock permeability and thickness. The aPC approach, which represents the models as a polynomial-based response surface, was then used to perform stochastic model reduction. Finally, a global sensitivity analysis was performed with Sobol indices based on the aPC technique to determine the relative importance of the different system parameters on pressure prediction. The research results indicate that there can be substantial uncertainty in pressure prediction locally, around the leakage zones, and that the degree of the uncertainty depends on the quality of the site-specific information available for analysis. The research results confirm the need for site-specific data for the efficient predictions of risks associated with geologic storage activities. Lastly, the research investigated the use of the AZMI model outputs as a basis for the Bayesian design of a monitoring framework for this zone of a geologic carbon storage system. Monitoring of reservoir leakage requires the capability to intercept and resolve the onset, location, and volume of leakage in a systematic and timely manner. The results of this research suggest that an optimal monitoring system with these capabilities can be designed based on the characterization of potential CO2 leakage scenarios that are determined using an assessment of the integrity and permeability of the caprock inferred from pressure measurements in the AZMI.