Assessment of Early Leak-Detection Capabilities at a Commercial-Scale Carbon Capture and Storage Site

Abstract:

In cooperation with the U.S. Department of Energy (DOE), a large Midwest carbon capture and storage (CCS) project will upgrade a power plant with oxy-combustion technology to capture approximately 1.1 million metric tons (MMT) of CO₂ each year. This project will design and construct a first-of-its-kind, near-zero emissions coal-fueled power plant that incorporates CCS. The project will implement a suite of monitoring technologies that includes early-leak-detection monitoring directly above the primary confining zone in regions of increased leakage potential (e.g., near wells that penetrate the caprock). In support of early leak-detection monitoring systems design, numerical models were developed and used to evaluate the relative value of various leak detection metrics over a range of hypothetical leakage scenarios. This preliminary modeling evaluation was based on a simplified model that assumed uniform properties for each model layer and interrogated both pressure and geochemical response in the first permeable interval overlying the primary confining zone. Simulation results indicate that pressure is likely to be the earliest indicator of leakage, given the rapid and areally extensive nature of this response. Simulated geochemical signals are much more localized and take much longer to develop than the pressure responses. Because of the buoyancy effect associated with supercritical CO₂ (scCO₂), early leak-detection monitoring for these leakage scenarios would be best achieved through monitoring in the upper portion of the interval near the contact with overlying low-permeability materials. Conversely, monitoring for geochemical signals associated with brine leakage exhibited less lateral spread than for scCO₂ cases and detection of leakage would be best achieved through monitoring at the base of the interval. Results from these preliminary models for a suite of leakage scenarios and monitoring location distances will be presented. These preliminary models will be updated as new site characterization data become available.