Micro-CT assisted NMR signal interpretation of mixed-wet heterogeneous carbonate at core scale resolution

Abstract

Limited work has been carried out to validate NMR interpretation on mixed-wet heterogeneous samples where conventional assumptions do not hold. Micro-CT image-based NMR simulation has evolved in the past decade allowing detailed investigation of NMR signal at controlled parameter observed in heterogeneous structure. Yet, no detailed work has been performed testing NMR signal response on tomographic images from real carbonate rock samples in the presence of surface relaxivity heterogeneity and microporosity that may cause diffusional coupling. This work uses series of micro-CT images of 1-inch bimodal carbonate sample at different saturation states accompanied by NMR measurements. The saturation states follow a typical workflow used in the industry for restoring samples to reservoir conditions. Two different conditions of wettability are considered: water-wet and mixed-wet. Given the series of tomograms for each wettability state, image porosity and saturations are derived via differential tomographic imaging. This spatial information is used in turn to carry out NMR simulations. A particular feature of the study is the treatment of micro-porosity. Uncertainty in physical properties associated with microporous regions such as effective diffusion and partial surface relaxivity was reduced by employing Archies Law and effective medium theory, assuming the tortuosity limit. These parameters were correlated to the intensity-based porosity. The effective medium approximation along with MICP measurements has been integrated to model the distribution of T2 relaxation times of the microporous regions for each fluid fraction at sub-voxel resolution.

The main focus of the study is to develop a workflow for integrating NMR measurements and micro-CT imaging approaches for deriving spatial distributed wettability models. Differential image technique, that tracks the relative water saturation change post spontaneous imbibition, was successfully employed in generating wettability information along the porosity label field. Two different spatial wettability models were used to simulate NMR responses which were compared to experiment measurement. The results were in acceptable agreement and show that the simulations are very sensitive to settings about diffusion coefficients in the microporosity, and that modelling the NMR response successfully can only be achieved if transport parameters are set correctly.