Enhanced recovery

Advances in Special-Core-Analysis Data Interpretations of Multiscale Measurements

The evaluation of complex carbonate reservoirs is a challenging task, and a detailed understanding of reservoir heterogeneity is still lacking. In this study, the effect of static rock type and sample size was investigated on special-core-analysis laboratory (SCAL) data derived from capillary pressure and resistivity-index experiments at reservoir temperature and net confining stress. Whole cores were generally found to yield different capillary pressure and saturation exponents that may not be possible to derive from average plug data.

Introduction

Most carbonate reservoirs are characterized by multiple-porosity systems that impart petrophysical heterogeneity to the gross reservoir interval. Reservoir heterogeneity is dependent upon the depositional environments and subsequent events in the history of the reservoir. In previous work, the effect of multiscale measurements (whole cores down to small trims) was studied on porosity/permeability and cementation factor. The whole-core samples were taken from a single well in a carbonate reservoir in Abu Dhabi, where 1.5-in. core plugs were drilled adjacent to the whole cores. Small trims were cut from the plugs for ­mercury-injection capillary pressure (MICP) determination. The main findings were that whole-core porosity matched log data very well and the plug porosities, when compared with the whole-core data, were systematically higher. The degree of heterogeneity played a major role in comparing whole-core to plug permeability: In the low-­permeability range (less than 10 md), whole cores could enhance 3D connection pathways and yield higher permeability values, whereas, in the high-permeability range, permeability can be overestimated in plugs through relatively larger porosity channels. The cementation factor was systematically reported to be lower in the measured whole-core samples. X-ray computed-tomography (CT) scanning was essential in explaining most of the variability seen in the measured rock properties at different scales. In another investigation on four different wells in the same reservoir, these earlier observations were confirmed on hundreds of multiscale samples.

The paper presents extended investigations from the same well to include porous-plate capillary pressure and resistivity-index measurements on individual whole-core samples and adjacent plugs. The measurements are conducted at reservoir temperature and net confining stress using dead crude oil and a simulated formation-brine-fluid system. The objective is to compare multiscale measurements to assess the effect of heterogeneity on Pc curves, saturation data, and the saturation exponent. Reservoir rock types (RRTs) on the core plugs were identified and discussed in a separate publication, where the effect of these rock types on residual-oil saturation and capillarity was studied thoroughly and found to be mainly dependent on the size range, the relative presence, and the distribution of pore throats. We discuss the effect of sample size from four different selected rock types in the reservoir. The whole-core primary-drainage capillary pressure curve seems to average the variability seen from core plugs. Whole-core resistivity-index measurements were on the lower-end of values from the plug measurements, especially when large variations were recorded from the plug data. For rock types with higher degrees of heterogeneity, whole-core imbibition capillary pressure data were found to deviate from individual-plug Pc curves and found to be a function of the relative presence of intermediate pore throats. These SCAL measurements from whole-core samples and from plugs provide the literature with high-quality multiscale measurements at representative reservoir conditions from different heterogeneous rock types that are rarely acquired. The experiments are the first of their kind in the industry and offer insight on the magnitude of possible variation between the whole-core and plug scales in such complex carbonates. They also confirmed that, with different scales, the influencing parameters stay the same, which in our case are the pore-throat sizes, their relative presence, and their distribution. In the future, we will acquire high-resolution X-ray CT imaging on the different rock-type samples at micro- and nanoscales to evaluate the effect of different pore types on the macroscopic measurements presented here.

Experimental Measurements

A fit-for-purpose conventional and ­special-core-analysis program was carried out on a carbonate core from the Middle East. The objective of the program was to measure macroscopic rock properties on various scales in order to identify the effect of heterogeneity on different volume measurements.

SCAL data from porous-plate capillary pressure and resistivity-index measurements are presented from four main RRTs in the reservoir. The measurements are acquired on 4-in.-diameter whole cores and 1.5-in.-diameter plugs. Helium porosity of the selected samples ranges from approximately 15 to 30%, and the permeability varies from approximately 5 to 1,000 md. One whole-core sample is selected to represent one rock type, including available core plugs from each rock type. Core plugs were cut adjacent to the whole cores as closely as possible. A clear trend is seen for porosity with rock type. Whole-core porosity is systematically lower than plug porosity, with one violation to this trend in RRT 4A depending on heterogeneity. Permeability presents less-­systematic variation with rock types and between whole cores and plugs. This is because of complex carbonate heterogeneity. It is interesting that the higher plug-permeability range yields lower whole-core permeability (see RRT 1A and RRT 3A). For the lower-range plug permeability (i.e., RRT 4A and RRT 3B), higher whole-core permeability values are obtained.

High-pressure MICP experiments were performed on end trims from each core plug to assist in rock typing. For the whole-core samples, top and bottom trims were cut from the same whole core to evaluate heterogeneity better and to have improved interpretations for any possible variation between whole-core and plug data. X-ray CT scanning and thin-section photomicrographs were also used in identifying variations in the internal structures and in rock typing.

Primary-drainage and imbibition experiments were performed on whole-core samples and plugs. All samples were thoroughly cleaned by flow-through techniques, using repeated cycles of several hot solvents to render the rocks water-wet (presumed wettability condition before oil entered the reservoir). The samples were saturated with 100% simulated formation water that is assumed to be representative of the reservoir-water composition. The capillary pressure and resistivity-index tests were conducted by the porous-plate method at net overburden stress and reservoir temperature by use of dead crude oil and simulated formation brine. During primary drainage, crude oil entered the rock samples at reservoir temperature. In this regard, aging would start as the oil penetrates the pore space. Several equilibrium capillary pressure steps were used in each saturation cycle, where the equilibrium criteria were no changes in saturation and resistivity at each step for at least 9 days (total duration at each Pc step was a minimum of 20 days). At the initial water saturation (Swi ) at the highest drainage, Pc aging continued for 1 month before the imbibition experiments were commenced.

Primary-Drainage Data

For RRT 1A, although the porosity/permeability crossplot shows variations between whole-core and plug data, the MICP curves obtained from plug trims and whole-core trims agree reasonably for the selected samples in this rock type. One may notice, however, slight variation between the whole-core MICP curves and the plug MICP data. This would be expected in such a heterogeneous carbonate reservoir. The porous-plate whole-core Pc curve agrees well with the corresponding MICP curves taken from the top and bottom of the whole-core sample. Similar observations can be derived for RRT 4A and RRT 3B. These observations may indicate that these rock types are less heterogeneous and apparently yield more or less similar macroscopic saturation measurements. Therefore, the heterogeneity effect in such rock types may have only minor effect on the multiscale macroscopic measurements.

For RRT 3A, however, there appears to be significant Pc variation within the plugs in the same rock type and within the whole-core sample itself. The whole-core top and bottom MICP curves are very different, indicating different rock properties along the whole-core length. This is rather confirmed by the X-ray CT slices from the whole-core sample depicted in Fig. 1. The top X-ray slice shows a darker gray, which indicates higher porosity and larger pores. This is indeed in line with the top MICP curve giving lower Swi than the MICP curve corresponding to the bottom section of the whole core. The porous-plate Pc curve from the whole core seems to average the rock-property variation seen in the plugs and within the whole-core sample itself. This can be viewed as an upscaled Pc curve from the various core plugs in this rock type, but directly measured on the whole-core sample. In this case, the whole-core measurement seems to be essential in capturing the heterogeneity in this rock type.

jpt-2013-8-specialcorefig1.jpg
Fig. 1—X-ray CT slices from the whole core in RRT 3A. The bottom slice shows a lighter gray color, indicating tighter rock nature. WC = whole core.

Whole-core capillary pressure curves and resistivity data from RRT 4A and RRT 3B seem to be averaging the slight variability seen in the corresponding plug data. The pore-throat-size-distribution curves from these two rock types show that they are less-heterogeneous rocks; nevertheless, the whole-core saturation-exponent values are measured to be at the lower range of the corresponding plug data. This observation is detected on all the involved rock types, a consideration of great importance for proper saturation determinations from resistivity logs.

Imbibition Data

All the plugs from RRT 3B gave similar imbibition behavior, which ties very well with the initially established static rock typing from petrophysical properties and geological descriptions. The whole-core sample shows behavior similar to that indicated by the plug imbibition curves, with small variation (higher entry pressure) that can be explained from the whole-core primary-drainage entry Pc curve being slightly higher than the plug curves. Another reason for the variation could be the very long time it takes for the whole-core sample to saturate, which could have resulted in nonequilibrium intermediate-saturation points during the imbibition process. Certainly, full equilibrium was attained at the lowest negative Pc, which gave ­residual-oil saturation (Sor) similar to that in the plug data.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 164140, “Advances in SCAL Data Interpretations of Multiscale Measurements From Different Carbonate Rock Types in a Giant Oil Field in Abu Dhabi,” by Moustafa R. Dernaika, Ingrain Inc.-Abu Dhabi; Samy Serag El Din and Zubair Kalam, ADCO; and Loay Hannon, Weatherford Laboratories, prepared for the 2013 SPE Middle East Oil and Gas Show and Exhibition, Manama, Bahrain, 10–13 March. The paper has not been peer reviewed.