Summary
A new technique for obtaining water-oil capillary pressure curves, based on nuclear magnetic resonance (NMR) imaging of the saturation distribution in flooded cores is presented.
In this technique, a steady-state fluid saturation profile is developed by flooding the core at a constant flow rate. At the steady-state situation where the saturation distribution no longer changes, the local pressure difference between the wetting and nonwetting phases represents the capillary pressure.
The saturation profile is measured using an NMR technique and for a drainage case, the pressure in the nonwetting phase is calculated numerically. This paper presents the NMR technique and the procedure for calculating the pressure distribution in the sample.
Inhomogeneous samples produce irregular saturation profiles, which may be interpreted in terms of variation in permeability, porosity, and capillary pressure.
Capillary pressure curves for North Sea chalk obtained by the new technique show good agreement with capillary pressure curves obtained by traditional techniques.
Introduction
Accurate petrophysical properties of reservoir rock such as capillary pressure, permeability, and relative permeability functions are essential as input for reliable oil in place estimations and for the prediction of the reservoir performance. Traditional methods for capillary pressure measurements are the mercury injection method, the diaphragm method, and the centrifuge method.
In the mercury injection method,1 the nonwetting phase is mercury which displaces a gas. The samples are usually evacuated to a low pressure and Hg is then injected in steps allowing for pressure equilibrium at each step, or alternatively Hg is continuously injected. Corresponding data on injected volume of Hg and the injection pressure are recorded. This technique is widely used for measuring capillary pressure functions for low permeability rocks. This is primarily because it is generally believed that pressure equilibrium in each pressure step is readily obtained, while this is normally a problem for other methods where a liquid is the wetting phase. The disadvantage of this technique is the uncertainty in the scaling of the measured data to reservoir fluid data and conditions.
In the diaphragm method or porous plate method, the problem concerning the scaling of the measured data is avoided, since this technique allows for the direct use of reservoir fluids. A water saturated sample is placed on a water-wet diaphragm to impose a boundary condition pc=0 to the wetting phase, i.e., the wetting phase is allowed to drain through the outlet end of the sample, at the same time as the nonwetting phase (oil or gas) is impeded. Pressure is added to the nonwetting phase and through a limited number of pressure steps, the capillary pressure curve is recorded. However, an important requirement is that equilibrium is obtained at each pressure step. This is the major problem when the diaphragm method is used on microporous materials. The drainage time may be considerable for each step, e.g., several weeks. In recent studies, thin micropore membranes have been used in an attempt to reduce the experimental time.2 Such a reduction will be less pronounced for low permeability rocks such as chalk since the flow resistance in the core is relatively more important.
In the centrifuge method, the amount of liquid produced from the outlet end of the plug sample at a certain spin rate is read directly from a measuring tube during rotation. From the geometry of the centrifuge, the spin rate and the average fluid saturation in the plug, it is possible to calculate the capillary pressure relative to the inlet end of the sample.3 However, a number of assumptions must be made3,4: the sample must be homogeneous and have a well-defined outlet pressure boundary condition, i.e., condition pc=0, and drainage equilibrium must be established at each spin rate. Most of these conditions can only be approximated in practice. For the centrifuge method, the condition of drainage equilibrium may be questionable even for sandstone samples.5 Slobod6 reported that equilibrium had not been attained for a 2 mD sample after 20 hr of spinning. King7 concluded that low permeability rock samples may suffer from very long equilibrium times. After 10 days of spinning in the centrifuge, a Berea sandstone sample of 200 mD had just reached equilibrium.
The objective of the development of the method presented here has been to avoid some of the disadvantages of the conventional methods described above. In this method a capillary pressure curve is obtained from a measured saturation profile after flooding the core. A similar experimental procedure was used by Richardson et al.8 to study end effects associated with flooding processes. The technique described here can be used with reservoir fluids. There is no porous plate to increase the flow resistance and the measurement of the capillary pressure function can be an integrated part of traditional flooding processes as performed with, e.g., unsteady-state relative permeability measurements. Only a very limited number of steps are needed, in principle only one step is required, therefore the time requirement for obtaining drainage equilibrium has not proved to be a problem. The technique utilizes the unavoidable end effect present in experiments with low permeability rocks. The capillary pressure function is obtained from the steady-state saturation profile in the core at drainage equilibrium.
