Detailed analyses of more than 50 core samples of western tight sands have resulted in several unanticipated observations that are set forth in this paper. Core analyses performed under stress paper. Core analyses performed under stress representative of producing conditions provided data on porosity, pore volume compressibility, stress dependence of permeability to gas, and slope of the Klinkenberg plot (permeability at constant net stress vs. the inverse of pore pressure). Scanning electron microscope (SEM) and petrographic microscope analyses were performed on samples cut from the ends of core plugs tested. The microscopic studies were explicitly plugs tested. The microscopic studies were explicitly directed toward observing the <0.1 micron flow path openings deduced from permeability data.
The Computer Operated Rock Analysis Laboratory(C.O.R.A.L.) used for,) these measurements has been previously described. Permeabilities are measured previously described. Permeabilities are measured with a maximum pressure drop of 20 psi, much less than the pore pressure of 100 to 1500 psia. At the one microdarcy level, the standard deviation of a sequence of permeability measurements under constant conditions is typically 2% of the measured value. Resolution is a few nanodarcies. The accuracy of porosity measurement is about +/− 2% of the reported value, but the sensitivity to pore volume change due to an incremental step in confining pressure is better than 0.1% of the pore volume. Thus, pore volume compressibility is measured to an accuracy of a few percent for a 1000 psi step in confining pressure. percent for a 1000 psi step in confining pressure. The selection of tight sandstone samples for analysis involved an intentional bias. Namely, all samples were from depths that were either known to be gas producers or judged likely to be producers on thebasis of wireline log analysis.
One of the sponsors of the work reported herein requested that studies be performed on western tight sand containing a broad spectrum of types and amounts of clays. To our surprise, the search for such samples a quite narrow range for both types of clays in the gas productive tight sandshis is illustrated by the data on five samples shown in Table 1. Although dry Klinkenberg permeabilities, under net stress representative of permeabilities, under net stress representative of producing conditions, varied by two orders of producing conditions, varied by two orders of magnitude, total clay content (<2 micron particles) of the samples was in the relatively narrow range of 3.8to 9.1 weight percent. The water-sensitive fraction of clays was found to be less than 50% of the total clay present and to lie in the range of I to 4 weight percent of the sample. percent of the sample. Neither the total quantity of clay nor the percentage of water-sensitive clay was found to percentage of water-sensitive clay was found to correlate with the porosity or permeability of the sample under pressures representative of producing conditions. however, the amounts of water-sensitive clay in the rocks were high enough for laboratory drying conditions to have a significant effect on the measured values of porosity and permeability.
Porosities and permeabilities for the samples in Table 1 were first measured after drying to constant weight at 60 deg. C at 45 deg. relative humidity as suggested by Bush and Jenkins). The measurements were then repeated after drying at the same temperature without humidity control. The percentage increases in porosity and permeability are shown in Table 2. The porosity and permeability are shown in Table 2. The increase in measured pore volume was examined in the context of the general rule of thumb by Bush and Jenkins 2 that "100 mg of water per gram of clay. equals one molecular layer of adsorbed water onmontmorillonite (smectite)." Assuming a density of 1.00 for relating this water to pore volume suggests that lack of humidity control resulted in driving about one layer of water of hydration off the expandable clay in each sample (see Table 2).
Klinkenberg permeability data, taken with net pressure on the core plug representative of the pressure on the core plug representative of the midpoint of reservoir drawdown, has been analyzed to deduce the size of flow paths. The analysis starts with the assumption that mass flow through a slot or narrow crack of uniform width can be described by the sum of Poiseuille's equation for laminar flow, with"no slip" at the walls, plus an empirical constant times Knudsen's equation for flow with a mean freepath larger than the opening. For a single slot of path larger than the opening. For a single slot of unit height this yields: