OIL SAND BASICS
Peter Pond called them "tar sands" in 1778 and in the early days of the oil business, tar sands were commonly called tar sands with a little bit of pride. The largest oil deposit in the world with a 400 year life span could not be sneered at. In today's politically-correct double-speak, we now call them "oil sands", not be be confused with conventional oil sands. So, at the suggestion of a good friend of mine, this page has been edited to remove the offensive word wherever possible.

 

Oil sands (tar sands, bitumen sands) are mined or depleted by steam assisted gravity drainage (SAGD) or in-situ 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.


Oil sands at 63 times magnification: shaly sand (left) with Vsh > 35%, clean sand (right) with Vsh < 5%.

The fluid column can be more complicated than conventional reservoirs. Here are some possibilities:
  1. bitumen with or without bottom water
  2. top water over bitumen with or without bottom water
  3. gas over bitumen with or without bottom water
  4. gas over top water over bitumen with or without bottom water
  5. any of the above with gas distributed unevenly in the main bitumen zone.

The oil sands of Alberta appear to be an easy task for a petrophysicist. After all, the sands are pretty clean, quite porous, and the fluid properties are reasonably well known. Even a novice geologist should be able to do it. However, a series of forensic log analyses over the last 30 years or so suggest that there are some basic misunderstandings about how oil sand cores are analyzed and how to calibrate log analysis results to that data.
 

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 Dean-Stark 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 Dean-Stark 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 Dean-Stark method, so the core porosity falls somewhere between total and effective porosity.

 

Dean-Stark core analysis (black dots) compared to total porosity
 (black curve) and effective porosity (left edge of red shading).

 

The calculated water saturation from Dean-Stark 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 Dean-Stark 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.

There are additional problems to resolve, as will be discussed below.
 

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.
 

DEAN-STARK CORE ANALYSIS METHOD
This method is used in poorly consolidated rocks such as oil sands and involves disaggregating the samples and weighing their constituent components. Samples are usually frozen or wrapped in plastic to preserve the contents during transport.

                                             Dean-Stark 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.
      1: WTwtr = WTsample - WTheated
      2: WToil = WTheated - WTminerals&clay

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.
      3: VOLwtr = WTwtr / DENSwtr / WTsample
      4: VOLtar = WTtar / DENStar / WTsample
      5: PHIcore = VOLwtr + VOLtar

Assuming clay bound water is driven off by heating and drying, then PHIcore equals total porosity. From comparison to log analysis results, it appears that some clay bound water remains in many cases, so PHIcore lies between total and effective porosity from log analysis.

Example of Dean-Stark porosity (black dots) showing that it is less than total porosity from
logs (black curve) due to incomplete drying of clay. Trying to match log porosity
directly to core may be futile in many cases. Porosity scale is 0.50 to 0.00.

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 Dean-Stark 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.

 



Dean-Stark core analysis in a water zone in an oil sand play (left side of table) contrasted with conventional helium porosity analysis .
 

OIL MASS FROM CORE LISTINGS
If not provided on the core listing, the equivalent value of oil mass from core analysis is derived from porosity, oil saturation, and an assumed oil density:
     1:  Woil = PHIcore * Soil * DENSoil
     2:  Wwtr =  PHIcore * Swtr * DENSwtr
     3:  Wrock = (1 – PHIcore) * GR_DENScore

Where:
  Soil = oil volume relative to pore volume
  Swtr = water volume relative to pore volume
  PHIcore = volume of water + volume of oil
  Woil = oil mass fraction
  Wwtr = water mass fraction
  Wrockcore = rock mass fraction

 

 

PHIcore

Star

Swtr

Vol Oil

Vol Wtr

GR_ DEN

WT Oil

WT Sand

WT Wtr

WT Rock

Oil Mass Wtar

Wtr Mass Wwtr

Rock ``Mass Wrock

frac

frac

frac

frac

frac

kg/m3

       

frac

frac

frac

0.306

0.301

0.699

0.092

0.214

2.650

0.092

1.839

0.212

2.143

0.043

0.099

0.858

0.271

0.236

0.764

0.064

0.207

2.650

0.064

1.932

0.207

2.203

0.029

0.094

0.877

0.279

0.306

0.694

0.085

0.194

2.650

0.085

1.911

0.193

2.189

0.039

0.088

0.873

0.244

0.304

0.696

0.074

0.170

2.650

0.074

2.003

0.168

2.246

0.033

0.075

0.892

0.298

0.217

0.783

0.065

0.233

2.650

0.065

1.860

0.233

2.158

0.030

0.108

0.862

0.273

0.298

0.702

0.081

0.192

2.650

0.081

1.927

0.191

2.199

0.037

0.087

0.876

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
If not provided on the core listing, the equivalent value of oil volumes from core analysis are derived from porosity, oil mass fraction, and an assumed oil density:
     1: Soil = Woil / (PHIcore * DENSoil)
     2: Swtr = Wwtr / (PHIcore * DENSwtr)
OR 2A: Swtr = 1.00 - Soil

Where:
  Soil = oil volume relative to pore volume
  Swtr = water volume relative to pore volume
  PHIcore = volume of water + volume of oil
  Woil = oil mass fraction
  Wwtr = water mass fraction

PHIcore

Star

Swtr

Vol Oil

Vol Wtr

GR_ DEN

WT Oil

WT Sand

WT Wtr

WT Rock

Oil Mass Wtar

Wtr Mass Wwtr

Rock Mass Wrock

frac

frac

frac

frac

frac

kg/m3

       

frac

frac

frac

0.306

0.301

0.699

0.092

0.214

2.650

0.092

1.839

0.212

2.143

0.043

0.099

0.858

0.271

0.236

0.764

0.064

0.207

2.650

0.064

1.932

0.207

2.203

0.029

0.094

0.877

0.279

0.306

0.694

0.085

0.194

2.650

0.085

1.911

0.193

2.189

0.039

0.088

0.873

0.244

0.304

0.696

0.074

0.170

2.650

0.074

2.003

0.168

2.246

0.033

0.075

0.892

0.298

0.217

0.783

0.065

0.233

2.650

0.065

1.860

0.233

2.158

0.030

0.108

0.862

0.273

0.298

0.702

0.081

0.192

2.650

0.081

1.927

0.191

2.199

0.037

0.087

0.876

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/01-quickmath.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 density-neutron 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 density-neutron 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 dual-water 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 semi-logarithmic 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.
      1: WToil  = (1 – Sw) * PHIe * DENSHY
      2: WTshl  = Vsh * DENSSH
      3: WTsnd = (1 - Vsh - PHIe) * DENSMA
      4: WTwtr = Sw * PHIe * DENSW
      5: WTrock = WToil + WTshl + WTsnd + WTwtr

Oil mass fraction:
        6: Woil = WToil / WTrock
        7: WT%oil = 100 * Woil

Typical densities are  DENSMA = 2650, DENSW = DENSHY = 1000, DENSSH = 2300 kg/m3.

Step 8: A bitumen pay flag is calculated with a log analysis oil mass fraction cutoff, usually between 0.050 and  0.085 oil mass fraction. A gas flag should also be shown on the depth plots where density neutron crossover occurs on the shale corrected log data.

 

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:
      1: OILtonnes = SUM (Woil * DENSoil * THICK) * AREA

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:
      2: OOIP = KV3 * SUM(PHIe * Soil * THICK) * AREA / Bo

Where:
  KV3 = 7758 bbl for English units    KV3 = 1.0 m3 for Metric units
  AREA = spacing unit or pool area (acres or square meters)
  Bo = oil volume factor (unitless)
  OOIP = oil in place as bitumen (bbl or m3)

 

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.

Water has a very high latent heat, so the volume of water to be steamed is as important to the economics as the volume of bitumen. High water saturation is bad news here, just as in conventional oil. Top water, top gas, and cap rock integrity are also major SAGD issues. The petrophysical analysis needs to look at the rocks well beyond the bitumen interval.


GAS EFFECT AT LOW PRESSURE
First lets look at the gas problem. If there is no gas crossover, you can skip this section. The conventional equation for porosity in a gas sand is:
     1: PHIe = ((PHInc^2 + PHIdc^2) / 2) ^ (1 / 2)

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:
     2: PHIe = ((PHInc^X + PHIdc^X) / 2) ^ (1 / X)

Where:
  X is in the range of 2.0 to 4.0, default = 3.0.
  PHIdc and PHInc are shale corrected values of density and neutron porosity respectively.

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.


PARTITIONING GAS and TAR VOLUMES
After shale volume and porosity have been calculated, water resistivity can be found in a bottom water zone below the oil, as these rarely has any residual oil. RW may vary somewhat in the oil sand interval and this can be adjusted if necessary by comparing calculated oil mass with core oil mass in non-gassy, relatively shale-free, intervals. Water saturation is then calculated from a shale corrected model such as Simandoux.

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
     4: Vhyd = PHIe * (1 – Sw)
     5: GasTarRatio = Max(0, Min((1 – OIL_MIN), (PHIDc – PHINc) / MAX_XOVER))
     6: Vgas = GasTarRatio * Vhyd
     7: Voil =  (1 – GasTarRatio) * Vhyd

Oil weight is calculated from log analysis as follows:
      8: WToil  = Voil * DENSHY
      9: WTshl   = Vsh * DENSSH

      10: WTsnd = (1 - Vsh - PHIe) * DENSMA
      11: WTwtr = Vwtr * DENSW
      12: WTrock = WToil + WTshl + WTsnd + WTwtr

Oil mass fraction:
      13: Woil = WToil / WTrock
      14:
WT%oil = 100 * Wtar

Where:
  OIL_MIN = minimum oil volume in gas zone as seen on core analysis, could be zero.
  MAX_XOVER =  maximum density neutron crossover in a gas zone (fractional)
  Vxxx = volume fraction of a component
  WTxxx = weight of a component (grams or Kg)
  Wxxx = mass fraction of a component
  WT%xxx = weight percent of a component

Comparison of oil mass from log analysis (solid line) with oil mass from Dean-
Stark core analysis (dots)  Oil mass scale is 0.30 to 0.00. Zone opposite this
caption is gas with residual oil; above and below are oil with no gas.

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:
     99:
Wtar = ((1.0 - Sw) * Phie * DENStar) / (DENSrna * (1.0 - Phie))

Here, DENSma is a computed result from the log analysis, and is usually wrong when gas is present. It hides the shale correction term and individual rock and fluid parameters cannot be adjusted. I strongly recommend that this "simplified" version be avoided.

It should be noted that core data is usually derived from a summation of fluids process, such as Dean-Stark 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, non-pay, 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.





"META/TAR" SPREADSHEET -- Log Analysis in Oil Sands
This spreadsheet provides a tool for Log Analysis of Oil Sands, including oil mass, net pay, and reserves calculations. 

Log Analysis for Oil Sands. English and Metric Units.



Sample of input data and crossplots for "META/TAR" Spreadsheet, used to analyze oil sand zones.




Sample of "META/TAR" net pay summary table.

 

TAR SAND EXAMPLE



Oil sand analysis with top water, bottom water, top gas, and mid zone gas. Core and log data match - but oil mass is the critical measure of success. Core porosity matches total porosity from logs, due to the nature of the summation of fluids method used in these unconsolidated sands. Minor coal streaks occur in this particular area.


   Water Satr'n
          Statistical

           Fluids
Model

      Fluids
 Determinist

    Oil Mass
ic Model

ALTERNATE MODELS

Comparison of petrophysical methods is often instructive. In the analysis shown at left, a probabilistic model (far left) is contrasted with a deterministic model (right). On the probabilistic model, oil is black, gas is red, water is blue, and clay bound water is gray. On the deterministic model, oil is red, gas is yellow, and water is white. Total porosity from core (black dots) and total porosity from log analysis are also shown on the deterministic model.

There are differences in porosity, especially in the low porosity range, differences in gas content, and differences in bulk water volume. Core oil mass (right) was used to calibrate the deterministic model; the match is excellent in both gassy and non-gassy oil intervals.

The statistical model was calibrated by comparing core water saturation to log analysis saturation (far left). The match is poor in some oil zones, reasonable in others, and of course is not a meaningful comparison in gassy zones.

The statistical model was tweaked several times but was never completely satisfying because the calibration to core was based on saturation and not on oil mass.

Oil mass comparison is the only correct way to match log analysis to core analysis in oil sand projects.

 

 



 


 

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