Inst. Fran9ais du Petrole; M.J. Argaud, SPE, Elf Aquitaine; and J.-P. Feraud, Total-CFP Summary.Laboratory equipment aimed at determining the exact correlation between resistivity and water saturation under stress, pressure, and temperature conditions is described in the first part of this paper. The porous-plate method adapted to reservoir conditions is used to obtain different saturation values during both drainage and imbibition. With this equipment, the influence of the effective stress on the porosity and formation resistivity factor can be studied before the test. In the second part of this paper, the values of the formation resistivity factor and resistivity index are compared for water-wet samples from sandstone and carbonate reservoirs. These measurements indicate that the influence of the effective stress depends on the nature of the rock sample. In addition, the resistivity/water-saturation law depends on the direction of the saturation change (drainage or imbibition) and on the nature of the fluids (water/oil or water/gas).

Experimental data of resistivity index and oil-brine capillary pressure on sandstone and carbonate rock samples from 4 reservoirs are reported. Laboratory equipment using actual fluids at reservoir conditions has been developed. The capillary drainage was achieved with the porous-plate method and fluid saturation along the core was checked for uniformity by conductivity measurements with a four-electrode system. For each rock sample, the resistivity index and capillary pressures were measured first with refined oil and then with crude oil. In addition, wettability indices were determined using Amott's tests. The resistivity/water-saturation law is well fitted by Archie's law except for vuggy carbonates. In this case, the resistivity/ water-saturation plot consisted in two straight line segments in log-log coordinates. The Archie's n-exponent values obtained with crude oil were different than those determined with refined oil except in one case. The n-value increases as the rock becomes more oil-wet. The maximum increase of n-value, from 1.68 to 2.19, was observed for one carbonate, initially neutral and becoming oil-wet during drainage with the Crude oil. Introduction The water saturation in oil reservoirs is generally estimated from resistivity well logs. The interpretation of these logs is based on two equations worked out by Archie: (1) (2) in which FR is the formation factor and IR is the electrical resistivity index. Equations (1) and (2) were determined for strongly water-wet clean formations (sandstone or unconsolidated sand with no clay). Here, a value of 2 for m and for n generally gives acceptable results for calculating the water saturation, Sw, determined by the following equation: S = [x] w resulting from the combining of Equations (1) and (2). The values of coefficients m and n are mainly obtained from laboratory tests performed on samples assumed to be representative of the reservoir to be investigated. But such tests are usually performed with simulated fluids and under ambient laboratory conditions, which differ greatly from reservoir conditions (pressure, temperature, effective stress, etc.). The values of m and n obtained by such tests and applied without correction to interpret resistivity logs sometimes lead to Sw values that contradict the ones obtained by other methods (preserved core analysis, etc.) or are incompatible with production observations. production observations. A great deal of research has been done on the influence of operating conditions (effective stress, temperature) on the measurement of the formation factor, hence on m. Reference 8 contains a detailed bibliographic study of this aspect. For nonclayey rocks, the influence of temperature is almost nil on the formation factor. In general, the increase in FR is greater than that linked solely to the reduction of. Hence m depends on the effective stress. The influence of operating conditions on the IR/Sw law and thus on n has received much less attention. This is probably linked to difficulties in performing truly representative laboratory experiments. As for m, it seems that the influence of temperature on n is slight or negligible, except for clayey rocks. The effective stress can cause an increase or decrease in the value of n. The level and direction of the variation of n depend on the type of rock (sandstone or carbonate) and on how the effective stress. is restored. Initial research on the influence of wettability on the resistivity/water-saturation law was done by making the pore surface entirely hydrophobic by a chemical treatment (Dri-film, silicone-containing products, etc.). Extremely high values of n (1.5 to 10 and sometimes even 20) measured under such conditions were strongly contested by log analysts because they led to resistivities of 10(6) to 10(9) oh m, which are values never encountered in reservoirs. P. 187

ABSTRACfAlthough a considerable amount of effort has been expended on the subject, the interpretation of shaly sandstone resistivities often proves problematic. Since the amount of high quality experimental data is somewhat limited, a joint research project was conducted. to acquire more laboratory data and to evaluate the data in terms of new and existing theoretical models.This project consisted of two sets of experiments:• conductivity measurements on fully brine saturated rock samples investigating the effect of changes in salinity and/or type of cations;• resistivity index measurements at decreasing water saturation and various salinities achieved by capillary drainage.The second set of experiments will be presented in this paper. Supporting data consist of mercury porosimetry, SEM analysis, and Qv as measured by chemical titration, membrane potential and C 0 vs Cw intercepts.Our data on resistivity index show that commonly used shalysand models such as Waxman-Smits or Dual-Water have difficulty predicting saturation when clay conductivity is dominant. This can occur when clay content is high or when brine conductivity is low. Models that incorporate additional geometrical information about the conducting paths, such as the Schwartz-Sen model or the DC model of Giouse and Argaud, do a better job of predicting the experimental results.