OIL SAND BASICS
Oil sands (tar sands, bitumen sands) are mined or depleted by steam assisted gravity drainage (SAGD) or insitu fire floods. In all these situations an adequate reservoir description is needed to assess the economics and progress of any project. The best oil sands are clean, medium to coarse grained, unconsolidated sands. However, they may be interbedded with finer, siltier, and shalier sands or overlain by lower quality reservoir rock. The log analysis needs to describe these variations, especially laterally continuous barriers to vertical flow of steam and oil movement.
The
fluid column can be more complicated than conventional
reservoirs. Here are some possibilities: In each case, the forensic analysis was undertaken at the request of a client who was unsatisfied with prior work that did not appear to provide an adequate description of the hydrocarbon potential in an oil sands reservoir.
Standard petrophysical analysis models are used for the volumetric determination of clay, porosity, water, and oil, and from this a realistic permeability estimate. Unfortunately, the DeanStark core analysis method, widely used to assess oil sand cores, does not measure volumes. Instead, the technique measures oil mass, water mass, and mineral mass. These are converted to mass fraction and then to calculated porosity and water saturation. Rarely, there may be some helium porosity and permeability data, but this is difficult in unconsolidated oil sands.
It is tempting to compare log analysis volumetrics to the DeanStark calculated volumetrics, and adjust log analysis parameters to obtain a “good match”. The biggest problem is that this form of core analysis gives a measure of porosity that is sometimes called “total porosity”, which includes clay bound water. In real life, some of the clay bound water is not driven off by the DeanStark method, so the core porosity falls somewhere between total and effective porosity.
DeanStark core analysis (black dots) compared to total porosity
The calculated water saturation from DeanStark also falls somewhere between total and effective, when some clay is present. Since log analysis gives effective porosity and saturation, we are comparing apples to aardvarks. The message is that log analysis cannot be calibrated directly to the core volumetric data when clay is present. Virtually all oil sands have some clay content somewhere in the interval of interest.
But we CAN calibrate to DeanStark core data in the mass fraction
domain, by converting the volumetric petrophysical analysis results
to mass fraction. That allows us to compare apples to apples, and
let the aardvarks go about their own business. Oil sand quality is
judged by its oil mass fraction and net pay is determined by an oil
mass fraction cutoff, not porosity and water saturation as in
conventional oil. So oil mass fraction is a mandatory output from a
petrophysical analysis. Oil in carbonates is also
extractable with SAGD, fire floods, or solvent floods. Gas is
usually less of an issue because there is less likelihood of
biogenic gas generation, but gas caps may exist in some plays.
DEANSTARK CORE ANALYSIS METHOD DeanStark laboratory apparatus In the lab, the still frozen cores are slabbed for photography and description, then samples are selected and weighed.
Samples are then heated and crumbled to drive off water, and
weighed again. The weight loss gives the water weight. Solvents
are used to remove oil. The sample is weighed again and
the weight loss is the weight of oil. The matrix rock is
separated into clay and mineral components by flotation, dried
and weighed again, giving the weight of clay and weight of the
mineral grains.
By dividing each weight by its respective density and
adjusting each result for the total weight of the sample, the
volume fraction of each is obtained. Porosity is the sum of
water plus oil volume fractions Because the bound water in
the clay is driven off by the drying sequences, this porosity is
the total porosity.
Example of DeanStark porosity (black dots) showing that it is
less than total porosity from If an oil sand is consolidated enough to be analyzed by conventional core analysis instead of Dean Stark methods (which can handle disaggregated samples), porosity, saturation , and permeability can be obtained. No permeability estimate can be made during a DeanStark analysis so permeability data in oil sands projects can be quite sparse. The table below shows a comparison of the results from both lab methods.
OIL MASS FROM CORE LISTINGS
Where:
Table 1 (above): When saturations and porosity are known (blue shading), all other terms can be calculated. GR_DENS must be measured or assumed and DENSwtr and DENStar are usually assumed to be 1000 Kg/m3. Some core analysis reports do the math for you, some do not. Since GR_DENScore represents a mixture of quartz and shale, this value should vary with shale volume. However shale volume is never reported on core analysis, so the composite grain density from the rock sample is used. If grain density is not recorded in the core analysis, we must assume a constant of 2650 Kg/m3 or lower.
FLUID VOLUMES FROM CORE LISTINGS
Where:
Table 2 (above): If oil mass fraction and water mass fraction are known, as well as core porosity (blue shading), all other terms can be calculated. Some core analysis reports do the math for you, some do not. OIL SAND MATH Petrophysical analysis of oil sands follows the standard methods that have been in use for more than 40 years: The math for these steps is at www.spec2000.net/01quickmath.htm , except where noted in the test. Step 1: Load, edit, and depth shift the full log suite, including resistivity, SP, GR, density, neutron, PE, caliper, and sonic, where available. If a thorium or uranium corrected GR (CGR) are available, load these too. Create a Bad Hole Flag if one is needed.
Step 2: Calculate clay volume. Because some uranium may cause spikes on the GR, use the minimum of the gamma ray and densityneutron separation methods. This eliminates false “shale” beds that would otherwise appear to act as baffles to the flow of steam or oil. The SP is unlikely to be a useful clay indicator due to the high resistivity of the oil zone.
Step 3: Calculate clay corrected porosity from the complex lithology densityneutron crossplot model. This model accounts for heavy minerals if any are present, compensates for small quantities of gas if present, and reduces statistical variations in the porosity values. DO NOT USE THE DENSITY POROSITY LOG ALONE. It will read too low if heavy minerals are present and too high if gas is present. The statistical variations at high porosity can give a noisy result. Some oil sands have enough coal or carbonaceous material to look like a coal bed. Set a coal trigger on the density and neutron and set porosity to zero when the trigger is turned on. There is nothing complex about the complex lithology model, so use it. See “Special Cases” below if there is gas crossover in the oil zone.
Step 4: Calculate clay corrected water saturation from the Simandoux or dualwater equations. These default to the Archie model in clean sands but give more oil in shaly sands.
Step 5: Correlate core porosity and core permeability on a semilogarithmic graph, if any data is available. The resulting equation takes the form Perm = 10^(A * PHIe + B) where A is the slope and B is the intercept at zero porosity on the graph.
Step 6: Calculate permeability as a continuous curve versus depth, using the regression analysis in Step 5.
Steps 1 through 6 cover the conventional volumetric analysis of an oil sand, but we are not finished yet.
Step 7: Convert log analysis
volumetrics to mass fraction values.
Typical densities are DENSMA = 2650, DENSW = DENSHY = 1000, DENSSH
= 2300 kg/m3.
Step 9: Oil in place is calculated from he standard volumetric
equation. However, some operators, especially surface mining
people, work in tonnes of oil in place. This equation is: Thickness is in meters and Area is in square meters.
If the oil equivalent in barrels or cubic meters is needed, the
standard equation can be used:
Recovery factor for surface mining operations is very high, maybe
0.98 or better. For SAGD, RF = 0.35 to 0.50 are used. Since we can't
keep the stream away from the shaly sands, recovery will vary with
the average rock quality in a SAGD project.
This equation is accurate enough for most gas zones,
but in very shallow gas sands, it will underestimate porosity. The
above equation must be replaced by:
Where: Density neutron crossover in a shallow gas sand with residual oil(shaded area) and core analysis porosity (dots). The low neutron porosity indicates little hydrogen content; the effect on the density is much smaller. An X of 3.0 or higher is needed to calculate effective porosity from logs. Porosity scale is 0.60 to 0.00 The exponent X is adjusted by trial and error until a good match to core porosity is obtained.
Many, but not all, gas zones related to oil sands have some residual oil. Hydrocarbon saturation is partitioned between bitumen and gas by the following method:
3: Vwtr =
PHIe * Sw
Oil weight is calculated from log analysis as follows: Where:
Comparison of
oil mass from log analysis (solid line) with oil mass from
Dean
Typical
densities are DENSMA = 2650, DENSW = DENSHY = 1000, DENSSH =
2300 kg/m3. This is the only way to rigourously calculate Oil Mass.
Other equations have been used, such as the one shown below, but are
less accurate, since shale volume is not explicitly enumerated: It should be noted that core data is usually derived from a summation of fluids process, such as DeanStark method, so the porosity from core matches total porosity better than effective porosity. Ditto water saturation. That's why we use oil mass and not porosity and saturation to calibrate log analysis to core data. Oil mass from log analysis is plotted, as shown at the right, along with oil mass calculated from core analysis data, on the depth plots to show the match between log analysis and core data results. The match between log analysis oil mass, porosity, and saturation with corresponding core data is usually excellent except in the very shaly, nonpay, intervals, mostly because the core data provided ignores shale and its effect on net grain density. The match in zones with high gas saturation varies in quality due to the inherent inaccuracy in the gas/oil partitioning calculation on the log analysis.
TAR SAND EXAMPLE


Page Views  Since 01 Jan 2015
Copyright 1978  2017 E. R. Crain, P.Eng. All Rights Reserved 