Ha'py field lies in the Mediterranean Sea off Egypt and is part of the Pliocene producing trend north of the Nile Delta. The original interpretation for the gas in place for Ha'py field by Amoco and others was 989 bcf. Re-evaluation of the field led to a gas in place estimate of 2,045 bcf.
Subsequent production in the field has substantiated the higher gas in place estimate under the re-evaluation.
The initial underestimation of the gas in place was a result of the use of an incorrect log-derived porosity model. This porosity model is not compatible with the petrophysical water saturation relationship used.
The incompatibility stems from the fact that petroleum engineers and petrophysicists have different definitions for the terms effective porosity and total porosity. The difference in definitions may lead a petrophysicist to improperly model the well logs to the available core analysis.
In addition, the engineer may mis-apply the petrophysicist's results to the reservoir characterization. This improper application leads to a substantial underestimation of hydrocarbon pore volume. The higher the clay and shale content in the reservoir, the greater the error.
We will clarify and explain definitions as they apply to core analysis, reservoir engineering, and petrophysics, and give examples using core analysis. Correct modeling of porosity data in these shaly reservoirs led to the higher estimation of the gas in place.
Introduction
Ha'py field is in the frontier development area of the outer shelf of the Nile Delta.1
The field, discovered by Amoco and International Egyptian Oil Co. (AGIP) in 1996, is one of the more significant gas discoveries in the Pliocene Trend, with gas in place estimated at 2,045 bcf (Fig. 1).
The field's main reservoir is the A20 sand of the Pliocene Kafr El Sheikh formation. In the Ha'py area, the Kafr El Sheikh consists of interbedded turbidite sands and prodelta shales.2 The A20 sand is an unconsolidated clay-rich prograding fan built of slump deposits. The A20 sand varies widely in thickness around the field, in places exceeding 100 m. The average porosity of the A20 sand is 30% with an average permeability of 10 md.
Structurally, Ha'py field lies within a province of northwest-southeast trending growth faults. The field is a structural trap formed where two of these growth faults merge. Kafr El Sheikh prodelta shales seal the trap, which is gas filled to the synclinal spill point. The reservoir ranges in depth from 1,350 to 1,750 m, a 400-m gas column.
Ha'py field was first recognized on the basis of a high-amplitude seismic anomaly, or "bright spot" (Fig. 2). Subsequent drilling of the Ha'py-1 discovery well confirmed the presence of gas in the A20 sand that created the amplitude anomaly.
Seismic mapping of the anomaly indicated a gas accumulation covering a large area (46 sq km, or 11,350 acres), but accurate determination of reserves within this area proved difficult. The anomalous interval exhibited variations in thickness and intensity, with numerous internal reflections (Figs. 2 and 3).
Furthermore, test results indicating a flow potential of up to 1.6 million cu m/day (56 MMcfd) appeared inconsistent with the shaly nature of the reservoir. To address the uncertainties in resource size and deliverability, we re-analyzed the data and conducted a reservoir characterization study.
Terminology
Petroleum engineers, in dealing with reservoir rocks, define two types of porosity: total porosity and effective porosity.
"Total porosity is the ratio of the total void space in the rock to the bulk volume of the rock; effective porosity is a ratio of the interconnected void space in the rock to the bulk volume, each expressed in percent."3
Conventional core analysis of clastic reservoirs, where the cores are cleaned and convection-dried in a vacuum oven at 105° C., yields total porosity. In clastic rocks, all pore space is interconnected, and thus, there is no difference between total porosity and effective porosity as given by the petroleum engineering definition above.4
As we shall see, shale can contain large amounts of total porosity. Log analysts have long wrestled with just how to deal with bound water. Whether or not to work with clay volumes (VCL) or shale volumes (VSH) has spurred many a debate.
The more enlightened petrophysicist defines effective porosity as "total porosity minus bulk volume clay bound water."5 In this definition, clay is defined as a "composition of crystalline fragments of minerals that are essentially hydrous aluminum silicates or occasionally hydrous magnesium silicates."6
Average shale content in most shales is between 30% and 55% bulk volume.4 7 Since shales are not all clay, they can contain significant amounts of effective porosity by the petrophysicist's definition. For example, Table 1 summarizes the average mineral content for both a sand and the adjoining shale in Seeligson field, South Texas.8 Since the average clay content of the shale is 38%, significant amounts of effective porosity can be expected.
Total porosity (convection dried) is relatively easy to measure on cores in the lab. Effective porosity, on the other hand is exceedingly difficult to measure in the lab. Effective porosity can be estimated by determining the clay bound water from the cation exchange capacity or by using air-brine and air-mercury capillary pressure data.5
Some analysts attempt to determine effective porosity by measuring porosity at 40% relative humidity and 60° C. The method attempts to leave several molecular layers of water on the clay that will approximate the volume of clay-bound water.9 The humidity-dried core porosity is greater than effective porosity but less than total porosity because some of the clay-bound water is extracted in the drying process.
Table 2 lists the results of core analysis using both the convection and humidity-dried methods. Note the large amounts of total porosity in the shales starting with sample No. 15. There is also a significant amount of humidity-dried porosity.
Table 3 compares effective porosity using air-brine and air-mercury capillary pressure measurements with convection and humidity-dried porosity. The data are from the same data set as the mineral content shown in Table 1. The air-mercury capillary pressure shale properties are similar to the shale properties in Ha'py field.
The problem
Fig. 4 displays a diagrammatic relationship between effective and total porosity, free and clay-bound water, and dry clay minerals as they relate to lithology varying from sand to shale.
The total porosity in the shale can be greater than, equal to, or less than the porosity in the adjacent sands. The difference generally depends upon the depth of burial and the effective stress. Over the years, log analysts have assumed that all water associated with shale was bound water (not a bad assumption in light of the fact that shales are the seals that prevent hydrocarbon migration).
Log analysts generally ignore the difference between clay and non-clay bound shale water and treat the determination of porosity and water saturation in distinctly different steps. When all of the porosity is filled with bound water, there is no effective porosity, thus no porosity available for the storage of hydrocarbons.
Log analysts have developed mathematical solutions that mimic the porosity reduction and force the effective porosity in shale to zero as the lithology grades from clean sand to 100% shale. When applied to a petrophysical interpretation, this misconception leads to a porosity that is composed entirely of movable fluids.
Gas is much more movable than water. High in the gas column above the gas-water contact, the only fluid present is gas. Extracting the bound fluid volume leaves only gas in the pore space.
The water saturation (Sw) approaches zero for the complete spectrum of shaly rock types. The Sw calculated though, after the bound water is subtracted, is the minimum original gas in place (OGIP) volumetric error.
The Sw error has a serious impact on reserve estimates. The underestimation of porosity by an amount equal to the Sw reduces the hydrocarbon pore volume. When the reservoir is shaly and unconsolidated, as in Ha'py field, the difference in hydrocarbon pore volume is quite substantial.
The reality of the situation is that shale contains large amounts of total porosity; both clay-bound porosity and significant amounts of effective porosity. In the original analysis of Ha'py field, all the water associated with the shale content was assumed to be bound.
The reduced effective porosity was calculated as a separate step. The reduced porosity raised the formation factor, which compounded the error and increased the water saturation significantly, instead of reducing the saturation as described above.
The Archie water saturation is not compatible with reduced porosity. The Archie a, m, and n values were determined using core porosity (total porosity). Thus an effective porosity that was too low resulted in a saturation that was too high (Table 4).
The problem created by assuming that all water in shale is clay-bound water is not new. Many attempts have been made to use effective porosity to calculate the correct core porosity. Most of these techniques involved shaping the shale content to increase log-computed effective porosity because the porosity is too low.
Some of these case studies are shown in Fig. 5. Larinov10 recognized that the amount of correction is related to the age of the formation; that older rocks needed less correction than younger rocks. The core-measured data support this concept as well.
Porosity from shale in older rock is typically less than that of younger rock due to more compaction and diagenesis. The porosity lines for core data with different total shale porosities are superimposed on the graph. One can see that both techniques can yield similar answers over limited portions of the lower shale volume range.
With the assumption that all water in shale is bound, the effective porosity is driven to zero at 100% shale, regardless of the age of the shale. In oil reservoirs, the shaliness where core porosity and effective porosity significantly diverge is generally of low enough permeability that oil trapped in the smaller pores will not produce.
If the correct shaping algorithm is used, the results may match production, even though the calculated oil in place is incorrect. Gas is different from oil, more movable, and emphasizes the inadequacy of the effective porosity approach.
Gas can and will produce at much higher shale volumes. A correct gas volume calculation is imperative to determine reservoir performance and economic viability. The difference between effective and total core porosity is not just semantics; it can represent large volumes of hydrocarbons.
Other core measurements use total core porosity, not log or core-derived effective porosity. Such measurements include capillary pressure and relative permeability. It is not correct to use these other measurements with effective porosity unless a correction is made to convert from one system to the other.
Fig. 6 illustrates the difference in Sw between a capillary pressure curve using total core porosity and effective porosity. Fig. 7 illustrates the difference in relative permeability curves between using total porosity and effective porosity. The difference is based on the fact that the hydrocarbon pore volume in both porosity systems must be equal.11
Petrophysical analysis
The responses of the neutron tool and the bulk density tool were modeled for various minerals and fluid contributions up to 50% porosity.12 Beyond 50% porosity, the models become unstable.
Fig. 8 shows the models for different rock and fluid combinations in Ha'py field based on quartz, gas (density = 0.115 g/cc, Ha'py 1 wireline data), water (density = 1.03 g/cc), and the clay minerals. The neutron and bulk density data for well Ha'py 1 is presented on the crossplot. The models calculated are the response lines and predicted porosities for the various minerals from 0% to 50% porosity.
From left to right, the magenta model is clean quartz and gas, no water. The next model, light green, is quartz, gas, and water at 50% water saturation. The blue curve is quartz and water, no gas. All three curves start at 0% sandstone porosity and extend to 50% porosity. The clay model, brown, represents the more common clay minerals found in the Ha'py reservoir.
A quick examination of the data relative to the models shows that almost all the reservoir appears to be in the range of 30% to 36% total porosity.
Also shown in Fig. 8 are the clay point used in the analysis and how the matrix point changes with increasing clay content. The brown line, from 0% sandstone porosity to 0% clay porosity, is the matrix line (0% total porosity line).
The analysis linearly combines the corresponding percentages of quartz and clay. When 50% clay is indicated, the matrix point for that rock is a linear combination of 50% sandstone and 50% clay. When 50% clay is modeled, the matrix density is 2.695 g/cc and the neutron value is 17.5 p.u.
These values represent 0% porosity for that rock type. If the rock has 50% total porosity, then the density value is 50% of the matrix density plus 50% of the fluid density. The neutron value at the matrix density is proportional to the fluid and rock properties as well.
Note that if the traditional approach is taken, using effective porosity and shales containing only bound water, the zero porosity line would be from the 0% quartz porosity point through shale points on the plot (southeast set of data points). Fig. 9 shows graphically the correction for gas in the computation.
The data for 50% clay and 30% porosity plot on the dashed line parallel to the gas-water line. All data on the black dashed line, if 50% clay, are 30% porosity. For the data to the right of the 50% clay matrix line, porosity is determined as if the rock were wet, no gas correction. Porosity to the right of the 50% clay matrix line is determined by following the line parallel to the sandstone-clay matrix line.
For the Ha'py well, in the interval from 1,550 m to 1,700 m TVD, the minimum total porosity in the shale is about 25%. All data parallel to the quartz-clay line will have the same porosity when 100% saturated with water. Following the line drawn at the minimum shale values for the neutron and density back to the intersection of the quartz-water line yields 25% porosity.
Fig. 10 from the Ha'py No. 2 well depicts a representative log calculation. Water saturation values from capillary pressure data were plotted on the log as purple diamonds. The water saturation curve in green was generated from the shale volume and the density porosity of the shale (volume of bound water divided by the total porosity).
The concept is that all the water in the sand is from the small pores, the shaly material that the gas could not invade. The shale-derived saturation matches the capillary pressure data very nicely, but neither matches the log-calculated saturation using a standard dual water technique (red curve).
A static downhole pressure for the field was determined after 570 days of production. The resultant P/Z plot (Fig. 11) suggests an interpolated OGIP of 2.32 tcf when the pressure equals to zero, in line with the calculated volume of 2,045 bcf from the re-analysis and reservoir characterization of Ha'py field.
Acknowledgments
We acknowledge the assistance of Simon Benavides in the preparation of this work. We also thank BP Egypt and Baker Atlas for permission to publish the material.
References
- Wigger, S., Bailey, J., and Larsen, M., "Ha'py Field: A Pliocene Bright Spot Example From the Nile Delta, Egypt," The Leading Edge, December 1997.
- Wigger, S., Simpson, M., Nada, H., Larsen, M., and Haggag, M., "Ha'py Field: The Result of Pliocene Exploration in the Ras El Barr Concession, Nile Delta," Egyptian Petroleum Conference 1996.
- Amyx, Bass, and Whiting, "Petroleum Reservoir Engineering," McGraw-Hill, New York, 1960, p. 39.
- Luffel, L., and Guidry, F.K., "New Core Analysis Methods for Measuring Reservoir Rock Properties of Devonian Shale," JPT, November 1992, p. 1,184.
- Hill, H.J., Shirley, O.J., and Klein, G.E., "Bound Water in Shaley Sands—Its Relationship to QV and Other Formation Properties," The Log Analyst, May/June 1979.
- American Geological Institute, "Dictionary of Geological Terms," Dolphin Books, Garden City, NY, 1962, p. 86.
- Pettijohn, F.J., "Sedimentary Rocks," Harper and Rowe, New York, 1957, 348 p.
- Truman, R.B., Howard, W.E., and Luffel, D.L., "Shale Porosity—Its Impact on Well Log Modelling and Interpretation," SPWLA 30th Annual Logging Symposium, June 11-14, 1989.
- Bush, D.C., and Jenkins, R.E., "Proper Hydration of Clays for Rock Property Determinations," Journal of Petroleum Technology, July 1970.
- Larinov, V.V., "Borehole Radiometry," Nedra, Moskwa, 1969.
- Juhaz, I., "The Central Role of Qv and Formation Water Salinity in the Evaluation of Shaly Formations," SPWLA Symposium 1979, Paper AA.
- Wiley, R., and Patchett, J.G., "CNL Neutron Modeling; A Step Forward," The Log Analyst, 1990, pp. 133-149.
The authors
William T. Bryant is a petrophysicist with BP PLC. He has over 30 years of industry experience in all aspects of formation evaluation. He is currently seconded to Kuwait Oil Co. by BP and is working in Kuwait.
Bob Truman is senior director of industry affairs for Baker Atlas. He has 37 years of formation evaluation experience in management, operations, research, and marketing. In his current position he is responsible for providing Baker Atlas with current and future industry outlook by interacting with leaders from client companies, universities, technical societies, government agencies, and industry analysts. He has a BS in mechanical engineering from California State University at Long Beach.
This was prepared for presentation at the SPE Annual Technical Conference and Exhibition in San Antonio Sept. 29-Oct. 2, 2002.
