SPWLA 46 th Annual Logging Symposium, June 26-29, 2005 NMR PETROPHYSICAL PREDICTIONS ON DIGITIZED CORE IMAGES C. H. Arns 1,* , A.P. Sheppard 1 , R. M. Sok 1 , and M.A. Knackstedt 1,2 1 Department of Applied Mathematics, Research School of Physical Sciences and Engineering, Australian National University, Canberra, Australia 2 School of Petroleum Engineering, University of New South Wales, Sydney, Australia * Corresponding Author: christoph.arns@anu.edu.au Copyright 2005, held jointly by the Society of Petrophysicists and Well Log An- alysts (SPWLA) and the submitting authors. This paper was prepared for presentation at the SPWLA 46 th Annual Logging Symposium held in New Orleans, Louisiana, United States, June 26-29, 2005. ABSTRACT NMR is a popular logging technique used to estimate pore size information, formation permeability, wettabil- ity and irreducible water saturation. Quantitative inter- pretation of NMR data is based on a set of fundamen- tal assumptions (e.g., pore isolation and fast diffusion). These assumptions establish the quantitative link between NMR response and petrophysical predictions. While there is a need to test these assumptions directly, to date no quantitative study on reservoir core material has been un- dertaken. The ability to digitally image reservoir rock in 3D, calculate petrophysical properties directly from the images coupled with a comprehensive simulation tool to numerically generate a range of NMR response data may help to address this need. In this paper we image a large set of reservoir cores in- cluding sandstones and carbonates at the pore scale using high resolution micro-CT. A set of petrophysical proper- ties are measured directly on the cores including surface- to-volume, permeability and pore size distribution. The permeabilities of the cores range from 10 mD to several Darcies. Realistic multiphase fluid distributions are de- rived by simulation of drainage. We then simulate the NMR responses on the same core images using a com- prehensive NMR simulator. The internal magnetic field is derived numerically from applied magnetic fields and susceptibility distributions and the phase evolution of the magnetic spins calculated with a random walk method. NMR responses currently include inversion recovery (T1) and CPMG (T2), and the longitudinal and transversal signals are monitored simultaneously. The interpretation of the signals acquired is done by standard 1D Laplace inversion to calculate the pore size distribution from T2 responses, and with a 2D inverse Laplace transform for fluid typing. In a preliminary study we compare predictions of petro- physical properties from the interpretation of the NMR response to direct calculations on the images. The foun- dational assumption of pore isolation is directly tested by partitioning of the pore space. This allows one to calcu- late the coupling constants and magnetisation exchange between pores, or between macro- and micro-porous re- gions. Further, the sensitivity of the responses to vari- ations in relaxivities or the presence of magnetic impu- rities is studied. Fluid typing is performed on a shaly sandstone sample. INTRODUCTION NMR techniques are usually employed in the petroleum industry to either predict permeability or for fluid typing. The former application uses the surface relaxation mech- anism or internal fields to derive a length scale (Brown- stein and Tarr, 1979; Kenyon et al., 1986; Kenyon et al., 1988; Kenyon, 1992; Song et al., 2000), which can then be used in permeability correlations (Banavar and Schwartz, 1987; Sen et al., 1990). The method relies on the application of the fast diffusion limit, in which the magnetisation of an isolated pore decays over time as a single exponential: (Wayne and Cotts, 1966; Brownstein and Tarr, 1979) M (t)= M 0 (t) exp  - t T 2  , (1) where M 0 is the initial magnetization and the transverse relaxation time T 2 is given by 1 T 2 = 1 T 2b + ρ S p V p , (2) with S p /V p the surface-to-pore-volume ratio of the pore space, T 2b the bulk relaxation time of the fluid that fills the pore space and ρ the surface relaxation strength. For small pores or large ρ the bulk relaxation contribution is considered negligible and 1 T 2 = 1 T 2s = ρ S p V p . (3) 1 MMM