Permeability Estimates from NML Measurements

Abstract

This method for estimating permeability from Nuclear Magnetism Log response incorporates the known manner in which NML response relates to pore size and to fluid and matrix properties. It also relies on the specific relations between mercury injection capillary pressure curves and permeability. Although the method as described here applies to sandstone, it is general enough for other applications. Introduction This paper presents a quantitative technique to estimate permeability from the Nuclear Magnetism Log response and from an independent determination of porosity. The introduction of new tools and porosity. The introduction of new tools and technology within the past decade has permitted lithology, fluid saturation, and porosity determinations from log response to become quite sophisticated. However, permeability estimation is still empirical and usually permeability estimation is still empirical and usually requires a detailed study of cores and logs to establish appropriate relations within a given field. The determination of permeability from logging measurements has remained elusive because the response of most logs is highly influenced by the composition, the fluid content and the electrical conductivity of the clays within the rock matrix. The Nuclear Magnetism Log (NML), however, responds only to unbound fluids within the rock pores. Since the size and configuration of that portion of the pore network containing unbound fluids also control rock permeability. a logical approach is to took to the NML as a means to deduce in-situ permeability. The NML has long been recognized as a means of conveying information related to permeability. Publications that appeared in the early 1960's, such Publications that appeared in the early 1960's, such as those by Hull and Coolidge, Brown and Ganison Coolidge and Gamson, and Wyman, all indicate that a qualitative relation between permeability and free fluid index (FFI) exists. Seevers subsequently developed an equation for sandstones, relating permeability to FFI, the observed thermal relaxation permeability to FFI, the observed thermal relaxation time of the fluid-saturated rock, and the thermal relaxation time of the saturating liquid when not confined in the rock pores. This equation required a constant of proportionality that remained constant within a given well but that varied from well to well. His method incorporated the physical model of Korringa et al. where the observed T1 was attributed to a component characteristic of bulk liquids and a shortened T1 component due to proximity to the rock matrix. It further depended upon a modification of the Kozeny equation, which related permeability to porosity, specific surface area, and tortuosity. Artus also considered an adaptation of the Kozeny equation and found that results, although encouraging, yielded calculated permeabilities that were consistently low. Timur et al. has used an adaptation of the equation proposed by Seevers to empirically relate permeability proposed by Seevers to empirically relate permeability of core samples to laboratory NMR measurements. All of these attempts to quantitatively relate permeability to NML response have depended upon some adaptation of the Kozeny equation. In addition, some are based upon a theory that does not take into account the diffusion of the protons within the liquid phase and the corresponding influence upon thermal relaxation time, T1. JPT P. 923