Abstract

A knowledge of how capillary pressure is related to saturation is important in predicting how one fluid immiscibly displaces another. Thus, there is a need for rapid, reliable methods for acquiring such data. This study investigates one such method, which makes use of a centrifuge. It differs from the conventional centrifuge method in that measurements are made on a long sample cut into slices. rather than on one short sample. Thus, it is possible to obtain simultaneously several points on the capillary pressure-average saturation curve. These data do not, however, represent the true capillary properties of the system under study. That is, they are affected by several experimental parameters, such as the length of the slices, and whether one or several speeds of rotation are used for obtaining a curve. These effects were studied in an attempt to get a better understanding of what the experimental data thus obtained really represent. Moreover, a new mathematical model, defined by four parameters, is proposed to correlate the data. It was found that the model parameters could be used to estimate reasonably well the area under a given curve, and, as a consequence, the model was considered to be satisfactory.

Introduction

A knowledge of how capillary pressure is related to saturation is important for two reasons. First, the relationship controls how fluids are distributed within the reservoir; and second, it provides, in immiscible displacement theory, the means of which the flow relationships for the displaced and displacing fluids are linked. In order to define this relationship, it is necessary to obtain accurate-experimental data.

Various techniques have been proposed for measuring capillary pressure. These include the restored state methodl, the mercury injection method 2, the dynamic method] and the centrifugal method44,5,6 The restored state technique is accurate, and it allows for the attainment of both the drainage and imbibition branches of the capillary pressure loop. However, the time required to obtain a curve is prohibitive. As a consequence, most of the data reported in the literature deal only with the drainage curve.

The mercury injection method, on the contrary, is a rapid method, but it makes use of mercury as the non-wetting phase, and of air-vapour of mercury as the wetting phase. It is, therefore, difficult to use the results obtained from this method to predict accurately the behaviour in an oil-water system. Moreover, the method is limited to the study of the drainage curve, and results in destruction of the samples.

In the dynamic method, both the wetting and non-wetting phases flow through the core, Consequently, it is more representative of actual displacement systems than static methods, and automatically corrects, at least partially. for such related factors as the difference between static and dynamic interfacial tensions and/or receding and advancing contact angles. However, it should be borne in mind chat this method, like all the others, requires that equilibrium be established before measurements can be made. Moreover, it is slaw and difficult to use.

This content is only available via PDF.
You can access this article if you purchase or spend a download.