Permeability estimation is crucial for formation evaluation, but faces challenges when used in low-permeability, unconventional formations. NMR well logging is often used to estimate formation permeability, but in many unconventional formations, the current NMR methods are not adequate.
We have developed a new method to estimate permeability using a modified Carman-Kozeny model with pore size, tortuosity, and porosity information inferred from NMR restricted diffusion measurements.
In this study, we focus on two low-permeability, organic-rich chalks (0.017 and 0.035 md) with connate water present. They are from the same formation but have different depths, TOC (total organic carbon), and bitumen content. These differences affect pore size, tortuosity, and permeability. The core samples are pressure saturated with two hydrocarbons—high-pressure methane or decane—with connate water present. NMR measurements are conducted under pressure to obtain the restricted diffusivity of the hydrocarbon-bearing pore space.
In planning the NMR restricted diffusivity measurements, an optimum series of diffusion-encoding times are chosen for the unipolar stimulated-echo pulse sequence to obtain the correlation between the restricted diffusivity (D) and free diffusion length (LD). By applying the Padé fit to the restricted diffusivity, we can better estimate the diffusive tortuosity (τ) and pore-body size (d) of the hydrocarbon-filled pore space.
The estimated pore-body size, tortuosity, and porosity from NMR are then used to predict permeability. We introduce a modified Carman-Kozeny model, which shows advantages over older methods like SDR and Timur-Coates models. The advantages of the new method are shown in organic-rich chalk with complex pore structures and organic matter.
This new method can potentially be used for estimating permeability by well-logging and core-log integration.