Prediction of gas–water relative permeability in tight rock from movable fluid distribution with nuclear magnetic resonance

Abstract

The measurement of the relative permeability in tight rock is challenging due to its ultralow permeability and the time-consuming nature of the experiments. Studying the movable and unmovable fluid distribution and establishing a reliable relative permeability prediction model is an urgent problem to be solved. This paper used nuclear magnetic resonance (NMR) to investigate movable and unmovable water distribution in tight sandstone under different centrifugal forces. A new method for predicting gas–water relative permeability in tight rock is established based on movable fluid distribution using the capillary bundle model. The results show that the distribution of movable and unmovable fluids is strongly influenced by the tight rock's pore size distribution and structure. The unmovable fluid saturation increases as the tight rock's permeability and median radius decrease. The nonlinear correlation between the NMR relaxation time and the pore throat size obtained from high-pressure mercury intrusion can be used to derive the pore size of the fluid distribution in tight rocks. The ratio of the movable fluid thickness to pore throat size increases near linearly with the logarithm of the pore throat size. The proposed mathematical model for the prediction of gas-water relative permeability based on movable fluid distribution is verified by comparing with the normalized relative permeability curve measured from experiments. This new model offers an alternative method of estimating the gas–water relative permeability when measurement is unavailable due to the ultralow permeability of the core samples.