Global Sensitivity Analysis of CO₂Circulated Geothermal Harvest from a Heterogeneous Reservoir

As geothermal harvest using geologically stored CO₂ is an emerging carbon utilization approach, coupled influences of formation heterogeneity, reservoir boundary, and well placement are still unclear. The aim of this study is to address this knowledge gap by a global sensitivity analysis using response surface method. Seven input parameters with specified ranges for both geostatistical and reservoir flow models form a 7-D parameter space, from which 100 input samples are drawn using Latin-Hypercube method. A set of CO₂ circulation model with heterogeneous reservoirs are setup using the 100 input samples and simulated. Four performance indicators are defined and derived from each of model simulations. Response Surface (RS) model for each indicator is trained based on the 100 input-indicator dataset. Global sensitivity indices are calculated using RS models, based on which sensitive input parameters for each indicator is identified for continuous sensitivity analysis in 2-D slices of RSs. It is found that the impact of heterogeneity is almost negligible to the four indicators, though it affects fluid and heat spatial distribution. Larger well space can increase lifespan, CO₂ storage, and total thermal recovery, but reduce power capacity. Higher injection overpressure can enhance power capacity, but shorten lifespan. Longer well perforation is beneficial to both power capacity and thermal recovery. Y boundary opening degree slightly and positively affects CO₂ storage only.