Experimental analysis of spatial correlation effects on capillary trapping of supercritical CO2 at the intermediate laboratory scale in heterogeneous porous media

Several numerical studies have demonstrated that the heterogeneous nature of typical sedimentary formations can favorably dampen the accumulation of mobile CO2 phase underneath the caprock. Core flooding experiments have also shown that contrasts in capillary entry pressure can lead to buildup of nonwetting fluid phase (NWP) at interfaces between facies. Explicit representation of geological heterogeneity at the intermediate (cm‐to‐m) scale is a powerful approach to identify the key mechanisms that control multiphase flow dynamics in porous media. The ability to carefully control flow regime and permeability contrast at a scale that is relevant to CO2 plume dynamics in saline formations offers valuable information to understand immiscible displacement processes and provides a benchmark for mathematical models. To provide insight into the impact of capillary heterogeneity on flow dynamics and trapping efficiency of supercritical CO2 under successive drainage and imbibition conditions, we present an experimental investigation conducted in a synthetic sand reservoir. By mimicking the interplay of governing forces at reservoir conditions via application of surrogate fluids, we performed three immiscible displacement experiments to observe the entrapment of NWP in heterogeneous porous media. Capillary trapping performance is evaluated for each scenario through spatial and temporal variations of NWP saturation; for this reason we adopted X‐ray attenuation to precisely measure phase saturation throughout the flow domain and apply spatial moment analysis. The sweeping performance of two different permeability fields with comparable variance but distinct spatial correlation was compared against a homogeneous base case with equivalent mean permeability by means of spatial moment analysis. Key Points: We investigate the impact of facies correlation on NWP saturation We conduct experiments with controlled heterogeneity and boundary conditions We evaluate Land model's ability to predict our experimental observations