Capillary Pressures and Gas Relative Permeabilities of Low-Permeability Sandstone

Abstract

Summary. Use of a high-speed centrifuge for capillary pressure measurements on low-permeability sandstones has been investigated. Curvatures for air/brine and air/decane displacements are compared. Capillary pressures of up to about 800 psi [5516 kPa] for gas displacing brine were extended over an order of magnitude by measurement of isotherms for desorption of water. Capillary pressures measured by high-speed centrifuge merged satisfactorily with results derived from the desorption isotherms. No distinct irreducible water saturation was observed. Permeabilities to gas at various levels of water saturation were also Permeabilities to gas at various levels of water saturation were also measured and found to be generally consistent with previous relative-permeability measurements. Introduction Low-permeability gas sands are a potentially abundant source of natural gas. Economical production of gas from the rock matrix generally requires that flow to the wellbore be aided by natural or induced fracture systems. However, production rate is probably still set by the gas flow rate from the matrix into the fracture system . This rate will depend on the detailed pore structure of the matrix and distribution of water contained therein. Capillary pressure measurements can be used to obtain information about pore size. Various possible methods are available by which capillary pressure measurements can be made on low-permeability gas sands, each of which has its own particular advantages and disadvantages. Mercury injection provides a convenient routine method. It appears to be a general characteristic of low-permeability gas sands, however, that water permeabilities may be up to one or two orders of magnitude less than Klinkenberg gas permeabilities. Because the nature of the wetting fluid has significant effects on pore structure and flow properties, capillary pressure determinations for which properties, capillary pressure determinations for which the wetting fluid is water or brine, rather than a vacuum as with mercury injection, are of interest. The high-speed Beckman L5–50P rock centrifuge appears especially suited for routine determinations of capillary pressures for gas/water displacements in tight sandstones. The centrifuge has a maximum speed of 20,000 rev/min; for water/air displacements, capillary pressures of up to 1,240 psi [8550 kPa] can be achieved. pressures of up to 1,240 psi [8550 kPa] can be achieved. The present investigation is focused mainly on the measurement, interpretation, and reproducibility of high-speed centrifuge data for tight sandstones. The samples studied were obtained from intervals of the Mesaverde formation during the course of the U.S. DOE Multiwell Experiment (MWX). Core properties are listed in Table 1. Summaries of the geologic setting from which these cores were taken are available. The high displacement pressures achieved by centrifuging are further extended by measurement of desorption isotherms to the stage where retention of water is mainly a result of surface adsorption. The complete capillary pressure curves obtained by combining centrifuge data and desorption isotherms are complemented by measurements of permeability to gas at various levels of water saturation and confining pressure. Experimental Methods High-Speed Centrifuge. Standard procedures recommended in Ref. 5 were used to determine capillary pressures for displacement of liquid by air with a centrifuge. pressures for displacement of liquid by air with a centrifuge. Core plugs 1 in. in diameter × 1 in. long [2.54×2.54 cm] were oven-dried at 158 deg. F [70 deg. C], evacuated, and then saturated with the liquid phase. which was either brine or decane. The brine contained 8% potassium nitrate with 200 ppm sodium azide added as a biocide. 8 Surface tensions of brine and decane measured at 77 deg. F [25 deg. C] are 71.2 and 23.7 mN/m, respectively-, densities were 1.044 and 0.73 g/cm, respectively. The initial speed of the centrifuge could be set at 2,000 rev/min without any liquid being expelled from the core samples. A small amount of silicon oil was included in the collection tubes to facilitate volume measurements for brine displacements; water was used for this purpose in air/decane displacements. Fluid volumes expelled were recorded as a function of time. The centrifuge speed was increased in increments of 500 rev/min up to 5,000 rev/min, and in 1,000-rev/min increments thereafter. Up to six core samples could be included in each experiment. Runs were terminated at 20,000 rev/min, the cutoff speed of the rotor. A typical experiment required 1 to 2 weeks for completion. Temperature is controlled by the evapatrol unit in the instrument that cools the vacuum chamber. The evapatrol setting was changed manually to maintain as constant a subambient temperature as possible. SPEFE P. 345