Abstract Archie[1] recognized the importance of information obtained over multiple observational scales. As confidence grows in properties measured by vastly improved logging capabilities, we tend to ignore Archie's original insight. Nuclear Magnetic Resonance,[2] NMR, yields rock independent estimates of porosity as well as estimates of bound and movable water and the distribution of pore bodies. These measures are combined with empirical core studies to infer permeabilities and to predict capillary pressure from NMR data. Our experimental study is designed to evaluate these empirical relationships and to understand the variability in T response due to the intrinsic properties of the porous network.
We have analyzed 90 cores with mixed orientation from five different clastic formations. The analysis included high pressure mercury injection, NMR desaturation, permeability, porosity and mineralogical measurements. NMR measurements were performed with a 2 MHz spectrometer. The porosity of the cores ranged from 4% to 23% while measured air permeabilities ranged from 0.01 to 900 md.
The T cutoff i.e., the boundary between free and bound water, for the complete set of samples ranges from 6 to 100 msec which represents significant departures from the typically assumed 33 ms cutoff for clastics. Mineralogical dependence is observed in the behavior of T cutoff. In general, the permeability estimation based on a weighted geometric mean of the T time performed better than the model based on the ratio of free fluid index to bound volume index. Mapping of pore bodies as measured with NMR to pore throat derived from Hg injection resulted in estimation of surface relaxivities which ranged from 6 to 50 µm/sec. These relaxivities were used to generate pseudo-capillary pressure curves from NMR which map well with the measured capillary pressures. Comparison between the cumulative NMR and mercury data yields insight into differences in the pore structure between samples. Mineralogical composition of the matrix influences the surface relaxivity especially if paramagnetic ions (e.g. Fe) are present.[3] On an average there is a general decrease in surface relaxivity with increase in quartz content.
Introduction New logging tools such as NMR are available to estimate or understand petrofabric. All scientific disciplines, e.g. medicine, physics, chemistry, geology and petroleum engineering have benefited from the application of NMR. NMR measures the net magnetization of the hydrogen atom (H) in the presence of the external magnetic field. The strength of the received signal is calibrated to the hydrogen atom concentration in a particular fluid. Porosity is then derived assuming that same fluid occupies the complete pore space. The fundamental equation governing the modern NMR relaxation spectra[4] is:
Equation (1)
The effect of the Bulk relaxation (T) is negligible and Diffusion relaxation (T) is considered to be zero as the experiments are performed under constant magnetic field. The two unknown parameter that control the surface relaxation are surface relaxivity (??, µm/sec) and surface to volume ratio (S/V) of the pore space. Both these parameters are difficult to measure directly but their influence on T measured by CPMG spin echo sequence is first order.
The main objective of this paper is to use NMR observations to infer and constrain pore geometries and petrophysical properties. Comprehensive suites of petrophysical measurement were made on 90 core samples from the five different clastic formations. All five formations have different depositional environments. Depositional environment controls the shape and size of the grains which in turn controls pore space geometry. The measurement suite includes porosity, permeability, mineralogy, NMR, capillary pressure and irreducible water saturation.