Leakage pathway estimation in geologic CO2 storage: application of inverse analysis to synthetic heterogeneous domains

Abstract

The purpose of this study is to examine the feasibility of inverse analysis to estimate a leakage pathway in a heterogeneous permeability field by using pressure anomalies from the intrusion of brine that is accompanied by gaseous CO2. To achieve this goal, the iTOUGH2 model is employed and applied to multiple brine reservoirs that are designed to realize pressure perturbations that are induced by leakage. Prior to inverse analysis, forward modeling with the TOUGH2 model illustrates brine and/or CO2 discharge through a leakage pathway that penetrates a caprock and the pressure anomalies in a permeable reservoir that overlies the caprock when CO2 is injected into a storage reservoir. The inverse modeling is applied to two types of heterogeneous permeability domains: (1) a renormalized model domain from the known heterogeneity and (2) a simplified model domain that consists of a few homogeneous sections with unknown heterogeneity. The inversion results from the simplified model domain are compared to those from the renormalized model domain to identify the applicability of the simplified model domain to detect a leakage pathway in complex heterogeneous permeability domains. This inverse modeling on the renormalized and simplified model domains examines the effect of the number of monitoring wells, monitoring periods and noises in measurements through 15 and 17 modeling scenarios, respectively. The accuracies of the inversions on the simplified model domain, which depend on the monitoring scenarios, are identical to those of the inverse modeling on the renormalized domain. Residual analysis illustrates that higher magnitude residuals in the storage formation should be weighted to match the residuals in the overlying formation to successfully estimate leakage pathways.