This work extends previously reported results to both higher and lower capillary pressures.
Abstract Recent experimental studies of the high-speed centrifuge technique have indicated that the method, if properly used, yields results which are in accordance with other methods. These studies have also shown that the concept of an irreducible wetting phase saturation is not valid. In the present paper work is reported in which these previously presented results are extended to both higher and lower capillary pressures. Air/liquid data have been obtained with n-decane as the wetting phase and are compared with new data (as well as with the previously reported data) for the air/brine system. The study involved 242 days of centrifugation and utilized 10 different Berea plug samples and 5 different high permeability plug samples taken from a producing reservoir.
The linear relationship between log Pc and log (or log Sw, as observed in previous work) was confirmed for Berea samples over the average saturation range from about 30 percent down to 10 percent. For lower average saturations, the saturation data are not sufficiently precise either to (1) demonstrate conclusively the persistence of this relationship, or (2) prove the breakdown of the Hassler boundary condition. For three of the five producing reservoir samples, it was found that two different linear segments characterized the log Pc versus log relationship, with. rather sharp transition occurring in the 30 to 40 percent average saturation range.
Introduction Capillary pressure measurements on core plug samples have long been used to determine water saturation as a function of height in petroleum reservoirs. Such data are important in evaluating original oil in place. Recent work, however, has indicated that both the determination and the interpretation of the data involve problems which are often overlooked.
These problems are of special concern in the range of low water saturation values. For example, a traditional assumption is that there exists a minimum value of the water or wetting phase saturation. Even for a fully water-wet situation, this assumption is now known to be incorrect. Unrecognized experimental difficulties in the use of the standard porous plate technique have clearly contributed to the unwarranted adoption of this assumption.
It should also be noted that various experimental techniques are now available for the determination of capillary pressure curves. Advances in most of these techniques have occurred in recent years. These advances have contributed to the recognition that significant problems relating to the measurement of capillary pressures still exist.
In the present paper attention is given to only one of the available methods - the high-speed centrifuge. Thus, the purpose of the paper is to report new capillary pressure data obtained by this technique and to discuss the implications of these results.
These new data involve saturation ranges and fluid systems for which there appear to be no previously published data. The data provide new evidence as to the nature of the capillary pressure versus saturation relationship in the low saturation region. Also, the data are used to evaluate possible limitations of the centrifuge method in this region.
Scope and Primary Objective In the present work the set of three Berea samples previously studied was used, as well as seven additional Berea samples with gas permeabilities ranging up to 1089 md. Also, five high permeability samples from a producing reservoir were used. Both air/brine and air/oil centrifuge runs were carried out. Capillary pressures ranged from 0.3 to 76 bars for the air/brine systems (six samples) and from 0.3 to 54 bars for the air/oil systems (six samples). In addition, air/oil capillary pressure data were obtained on three samples over the range from 0.07 bars to 8.5 bars. These samples included one Berea sample and two producing reservoir samples. In Table 1 the permeabilities of the various samples, the fluid systems and the maximum capillary pressure values reached in the present work are listed. Mineralogical data for these samples will be presented in a forthcoming paper.
One of the objectives of the work was to test the prediction that at sufficiently high capillary pressures the assumed Hassler boundary condition may fail. According to this assumption, the bottom end face saturation of core plug in a high-speed centrifuge is always maintained at 100 percent. Two recent studies were in large part designed to provide evidence relating to the possible failure of the Hassler assumption. However, neither study appears to have provided conclusive evidence on this point. In neither case was the maximum centrifuge speed sufficiently high that the expected critical capillary pressure was substantially exceeded.
The final column of Table 1 records the values of the expected critical capillary pressures for the various samples and centrifuge conditions used in the present work. These critical values were calculated from an expression given previously. The permeabilities used in these calculations were liquid permeabilities measured just prior to the particular centrifuge run shown in Table 1. Surface tension values were taken from the literature. It is seen from Table 1 that the expected critical capillary pressure was exceeded by more than 50 percent in only one of the runs carried out in the present study (Run 4). Thus, during the first three runs, the actual critical capillary pressure mayor may not have been reached at the highest centrifuge speed used.
Another feature of the work is the very long run times used at the two highest speeds in the first four centrifuge runs listed in Table 1. The average run times for these speeds were 8 and 12 days, respectively. In all, the study reported here involved a total of 726 sample-days of centrifuge operation. These long run times were considered to be important in testing the possibility of a failure of the Hassler boundary condition.