Interpreting Permeability From Mercury Injection Capillary Pressure Data

Abstract

Laminar flow theory predicts a strong correlation between permeability and pore-throat distribution as revealed by Mercury Injection Capillary Pressure (MICP) data. Previous studies have developed relationships between MICP data and permeability; however, the permeabilities predicted by different methods can differ substantially from the measured permeabilities and from each other, especially in low permeability samples of interest for unconventional reservoirs. The purposes of this study are to evaluate why there is such large scatter, identify algorithms that best predict permeability over a wide range of permeabilities, and evaluate what type of permeability is actually measured by MICP data. Precision of permeability predictions is low due to insufficient MICP pressure measurements, assumption of MICP curve shape, permeability anisotropy of geological samples, and low precision and accuracy of permeability measurement of tight rocks. Four methods for estimating permeability from MICP data are found to have small bias and reasonable precision over a wide range of permeability: the modified Purcell, the Katz-Thompson Lc, Katz-Thompson Lh, and the Swanson methods. A weighted average of these permeability estimates corrects for accuracy problems and increases permeability estimate precision. However, this MICP-predicted average permeability still varies from measured Klinkenberg-corrected steady permeability by an average of a factor of 2. This mismatch may be more apparent than real. Restoring reservoir stress prior to conventional permeability measurement fails to remove completely the core damage caused by microfractures created during extraction, preparation, and storage of tight rock samples from deep boreholes. MICP permeabilities are estimated from the pore-throat distributions, which do not include the significant flow contributions from microfractures. Difference between MICP permeability estimates and measured permeability of tight samples may be caused by the inability of conventional permeability analysis to remove damage effects by stress restoration. If so, MICP permeability estimates are as good as or better than permeability measured from tight, subsurface samples. MICP permeability is either the ambient matrix permeability or a stressed matrix permeability, depending on the relative magnitude of in situ reservoir stress and Hg pressure at threshold saturation.

AAPG Datapages/Search and Discovery Article #90216 ©2015 AAPG Annual Convention and Exhibition, Denver, CO., May 31 - June 3, 2015