John H. Dunsmuir

Pore and throat size distributions measured from synchrotron X-ray tomographic images of Fontainebleau sandstones

The three-dimensional geometry and connectivity of pore space controls the hydraulic transport behavior of crustal rocks. We report on direct measurement of flow-relevant geometrical properties of the void space in a suite of four samples of Fontainebleau sandstone ranging from 7.5 to 22% porosity. The measurements are obtained from computer analysis of three-dimensional, synchrotron X-ray computed microtomographic images. We present measured distributions of coordination number, channel length, throat size, and pore volume and of correlations between throat size/pore volume and nearest-neighbor pore volume/pore volume determined for these samples. In order to deal with the ambiguity of where a nodal pore ends and a channel begins, we apportion the void space volume solely among nodal pores, with the channel throat surfaces providing the nodal pore delineations. Pore channels thus have length but no associated volume; channel length is defined by nodal pore center to nodal pore center distance. For a sample of given porosity our measurements show that the pore coordination number and throat area are exponentially distributed, whereas the channel length and nodal pore volume follow gamma and lognormal distributions, respectively. Our data indicate an overall increase in coordination number and shortening of pore channel length with increasing porosity. The average coordination number ranges from 3.4 to 3.8; the average channel length ranges from 200 to 130 μm. Average throat area increases from 1600 to 2200 μm 2 with increasing porosity, while average pore volume remains essentially unchanged at around 0.0004 mm 3 .

Morphology and physical properties of Fontainebleau sandstone via a tomographic analysis

We present a study of the morphology and bulk physical properties of a Fontainebleau sandstone via an X ray tomographic analysis. Synchrotron-based X ray tomographic techniques provide us with a high-resolution (7.5 μm), three-dimensional digitized representation of the sandstone that leaves the sample intact and unaltered. To estimate a wide spectrum of bulk properties of the Fontainebleau sandstone specimen, we extract from this image a number of different correlation functions that statistically characterize the pore-space morphology and relevant pore-space length and time scales. These statistical measures are obtainable from lineal, plane, and/or volume measurements and include the porosity, specific surface, two-point and three-point probability functions, lineal-path function, chord-length distribution function, pore-size distribution function, and coarseness. The pore-size distribution function, in particular, contains a certain level of connectedness information and accordingly can only be obtained from a three-dimensional representation of the sample. Many bulk properties of the sandstone, such as the mean survival time τ (obtainable from Nuclear Magnetic Resonance relaxation studies), fluid permeability k, effective electrical and thermal conductivities, and effective elastic moduli, can be estimated using the aforementioned statistical correlation functions. Specifically, the electrical conductivity (or, equivalently, the formation factor F), mean survival time, and fluid permeability are determined using rigorous bounds. The mean survival time and fluid permeability are also found using direct simulation techniques and cross-property relations, respectively. One such cross-property relation for k depending on τ and F gives a permeability estimate that is within a factor of 2 of the experimental result.

Use of X-ray computed microtomography to understand why gels reduce relative permeability to water more than that to oil

Abstract X-ray computed microtomography (XMT) was used to investigate why gels reduce relative permeability to water more than that to oil in strongly water-wet Berea sandstone. XMT allows saturation differences to be monitored for individual pores during various stages of oil, water, and gelant flooding. The method also characterizes distributions of pore size, aspect ratio, and coordination number for the porous media. We studied a Cr(III) acetate–HPAM gel that reduced permeability to water (at Sor) by a factor 80–90 times more than that to oil (at Swr). In Berea, the gel caused disproportionate permeability reduction by trapping substantial volumes of oil that remained immobile during water flooding (i.e., 43.5% Sor before gel placement versus 78.7% Sor after gel placement). With this high trapped oil saturation, water was forced to flow through narrow films, through the smallest pores, and through the gel itself. In contrast, during oil flooding, oil pathways remained relatively free from constriction by the gel.

Irreducible water distribution in sandstone rock: Two phase flow simulations in CT-based pore network

Abstract A simple cutoff approach based on the capillary bundle model has become an industrial standard method to quickly obtain the irreducible water saturation from NMR T 2 distribution. However this approach is not always valid. To overcome the shortcoming of the capillary bundle model that ignores pore-pore connectivity, we have conducted a two-phase flow simulation in a CT-based pore network. The CT -based pore network is a representation of a real rock pore structure that is described by a binary X-ray tomographic data set. Simulation on such network mimics a process of the porous plate measurement. The generated capillary curve is quite reasonable. The oil accession distributions at different water saturations plotted as a function of pore size provide an insight for the immiscible displacement process in the real rock pore structure. Water is trapped not only in dead ends but also in the well-connected pores due to a pore level by-pass mechanism. At the capillary end point pressure, a plot of the trapped water distribution as a function of pore size has the same lognormal distribution as the pore size distribution, which is much different from what the simple capillary bundle model suggests.

Local Reconstruction Applied to X-Ray Microtomography

A characteristic of local reconstruction is that the x-ray detector array need only be slightly larger than the projected volume of interest within the object scanned. This feature of local reconstruction has several practical consequences for to- mographic imaging. We present an example in which local reconstruction extends the capability of a micro-CT scanner beyond the limits set by those physical characteristics of the scanner components that would be needed if standard global tomographic reconstruction algorithms were to be used.

NMR T2 distributions and two phase flow simulations from x-ray micro-tomography images of sandstones

The distribution of fluids in the pore space of a series of sandstones is calculated as a function of capillary pressure using a two phase flow simulation model. The pore space is represented by a system of channels and nodes which are derived from x-ray micro-tomography images of sandstones. The sandstones studied varied in permeability from approximately 40 to 3,000 mD. The simulation results illustrate the significance of the pore level by-pass phenomena in controlling the location of fluids within the pore structure. The implications of these results on the interpretation of NMR T(2) distributions to determine the irreducible water saturation are discussed.

Electric field driven flow in natural porous media.

Electric fields were applied to fluid-saturated packed sand beds (0.23+/-0.03 mm average pore diameter), and the effects on the mobility of the water molecules were monitored using stimulated echo (STE) and pulsed field gradient (PFG) experiments. The mean flow velocity, averaged over the entire sample, is expected to vanish in closed systems, but the PFG and time dependent signal decay was enhanced beyond the effects of thermal diffusion, due to velocity dispersion. The internal flow generated by the electric field was shown to be fully time-reversible upon inverting the electric field polarity (for total flow times of up to 0.4s), a strong indication that the NMR detected displacements were mainly due to electro-osmotic flow (EOF). However, a comparison of the velocity dispersion for different electrolyte concentrations showed that the measured effect scaled with the applied power VI (V = voltage, I = electric current), rather than with the voltage alone, contrary to the prediction of the basic model for EOF in a single capillary channel.

Effective medium models from x‐ray microtomography images.

A novel method is described where an effective medium model is generated from a 3‐D x‐ray microtomography image of a rock sample. In contrast to current modeling practice, material phases are not assigned idealized geometries, like spheroids. Instead, strain concentration tensors are computed numerically for the true phase geometries observed in the 3‐D image. In this method, a 3‐D image of x‐ray attenuation is converted into a 3‐D elastic properties volume, the strain field is computed within the volume, and the strain concentration tensors, which quantify the contribution of each grain and pore type to the bulk elastic properties, are calculated from the strain field. From only one representative sample of a rock type and with no assumptions about grain or pore shape, an effective medium model is generated that specifies the effective elastic tensor as a function of the various mineral and pore volume fractions. The method is demonstrated by generating effective medium models for several rock types. The...