An optimization approach to design monitoring schemes for CO2 leakage detection

This paper demonstrates an approach to identify optimal monitoring designs that minimize the time to first detection of CO2 leakage from a subsurface storage formation. This research is part of the National Risk Assessment Partnership (NRAP), a DOE project tasked with conducting risk and uncertainty analysis in the areas of reservoir performance, natural leakage pathways, wellbore integrity, groundwater protection, monitoring, and systems level modeling. Our approach applies a simulated annealing algorithm that searches the solution space by iteratively mutating potential monitoring designs. A user-defined leakage signature based on change from initial conditions is implemented to infer CO2 leakage had occurred. An example application was performed to demonstrate the effectiveness and efficiency of this method compared to an exhaustive search of the entire solution space. We accounted for uncertainty in the example by evaluating the performance of potential monitoring designs across a set of simulated leakage realizations. The example application approached the optimal time to leakage detection in a few minutes using a standard workstation, which was several orders of magnitude faster than an exhaustive search of the solution space, thus exhibiting efficiency and effectiveness. A user-friendly tool, DREAM (Designs for Risk Evaluation and Management), is being developed for use on personal computers to make this method accessible to stakeholders, regulators, and researchers.