CRAIN'S
PETROPHYSICAL
HANDBOOK
c. 1978 - 2008 E. R. (Ross) Crain, P.Eng.
Rocky Mountain House, Alberta Canada T4T 2A2
403-845-2527 email us
Please be fair to the author - pay your Shareware Fee HERE
and receive a copy of Crain’s Petrophysical Handbook on CD-ROM
at no extra cost.
Updated 30 June 2005
CHAPTER NINE: CALCULATING LITHOLOGY Table
Of Contents
SPECIAL
FEATURE: Guest Chapters by Dr Zoltan Barlai Publication history: This Chapter formed Chapter Nine of The Log Analysis Handbook published by Pennwell (1986). Extensive revisions were added for this electronic edition in June 2002 (Sections 9.08 through 9.12, 9.15). Section 9.00 updated Oct 2003, Jan 2008. Section 9.16 and 9.17 added Oct 2004, updated Jan 2008.. CHAPTER NINE: CALCULATING LITHOLOGY 9.00
Introduction To This Chapter In oil field applications of logs, interest is primarily directed to definition of the amount and type of fluids in the formations. These determinations require that matrix effects be defined and accounted for through appropriate assumptions about the mineralogy of the reservoir or by combinations of logging measurements that automatically compensate for mineral effects. In addition, we have found that a knowledge of the mineral composition of the reservoir aids in understanding its depositional environment, porosity distribution, production characteristics, and exploitation potential. In coal, evaporite, and mineral exploration, the primary interest is in the identification of the minerals - porosity is usually negligible. Mathematically, the oil-field and mining situations are identical, so the methods described here apply to both disciplines. Before proceeding, we need to define the nature of rocks more clearly. An element is a primary component of a chemical compound. Familiar elements are iron (Fe), calcium (Ca), carbon (C), and oxygen (O). A mineral is a naturally occurring inorganic compound with a specific chemical formula and a defined crystal structure. Many naturally occurring minerals are impure, so their chemical makeup varies slightly. Familiar mineral compounds are quartz SiO2) and calcite (CaCO3). A rock is made from a mixture of minerals, although one mineral may dominate. For example, some sandstones are mostly quartz (SiO2) but many other minerals may also be present. Other sandstones may be mostly feldspar with little quartz. Limestone is a rock containing mostly calcite (CaCO3) but other minerals may be mixed with it. Most rocks have a wide range of minerals and the fraction of each mineral in a rock may vary widely from one sample to another. Minerals are often described by, and hand samples identified by, their hardness, magnetic response, colour, luster, streak, cleavage, crystal form, specific gravity, reaction to acid, or even their taste and smell. These terms are useless for petrophysical log analysis, which relies on physical properties that can be measured remotely in a well bore, such as density, acoustic velocity, neutron and gamma ray response, or electrical resistivity. Minerals are classified into groups and sub-groups. Silicate minerals have silicon and oxygen in their composition. There are four types of silicate minerals:
Silicates are also divided into two groups based on their color and density. Light (nonferromangesian) silicates are light in color and have a specific gravity around 2.7. Light silicates contain various amounts of aluminum, potassium, calcium and sodium. Dark (ferromagnesian) silicates are dark in color and have a specific gravity ranging from about 3.2 to 3.6. They contain mostly iron and magnesium. All other minerals are put into the non-silicate group, then broken down into six subgroups:
Because the earth has an active surface, minerals (in the form of rocks) are under constant change. Molten rock from the interior of the earth can be exposed at the surface from volcanos or mid-ocean ridges. When molten these rocks are called lava flows and when cool they are called igneous rocks. As igneous rocks are eroded by weather and water, they become lose grains or dissolved in water. When deposited, they become soil or sediment, and later under the pressure of overburden, turn into sedimentary rocks. If sedimentary rocks are forced deep enough, heat and pressure modify the rock structure. These are called metamorphic rocks. All three kinds of rocks can contain porosity that can hold economic quantities of oil and gas, although sedimentary reservoirs are much more common. Any of these rock types can re-enter the mantle and become molten again, by subduction at the edges of tectonic plates. This cycle of igneous – sedimentary – metamorphic is called the rock cycle. The majority of this Chapter deals with sedimentary rocks because the majority of petrophysical analysis is performed on sedimentary rocks. However, igneous and metamorphic rocks do form important reservoirs. See Section 9 .17. Sedimentary rocks are an accumulation of fragments of many pre-existing rocks. Weathering is a process by which rocks are broken down into sediments. There are two types of weathering:
Transport describes the process by which sediments are moved across the surface. Types of transport include fluvial, glaciers, wind (aolean), and gravity. Depositional environments describes where sediment comes to rest, The three main groups however are:
Lithification is the process by which sediments come together to form a sedimentary rock. There are three ways in which this is done:
The texture of a rock is based on the size, shape, and arrangement of the grains and other parts of the rock. Sedimentary rocks can be broken down into five different textures:
Mineral composition in sedimentary rocks varies widely.
Many descriptive terms are used to define rock samples, most cannot be determined directly from petrophysical logs. Shape, sorting, bedding type and bed thickness are common terms. Size of the sedimentary particles is a semi-quantitative approach to sample description and assists the petrophysicist in understanding the rock texture. Terms used are:
The kinds of rocks we can identify with well logs depend on the logging tools that have been run in the well bore, the rock mixtures present, and local zone knowledge. In clastic and carbonate sections, we can usually identify quartz, shale, limestone, dolomite, anhydrite, coal, pyrite or glauconite or siderite or other heavy minerals, salt, potash, trona, sulphur, gypsum, and a few rarer minerals like fluorite or barite, provided the minerals occur as mixtures of only a few components and we have a full modern log suite. Shale minerals, such as montmorillonite, illite, and chlorite, can be distinguished if we have additional logs. Kaolinite and feldspars can also be defined under certain conditions, as can mica. Although not discussed in this Chapter, hardrock minerals and uranium deposits can be evaluated with well logs. The mineralogy of unconventional reservoir rocks, such as granite, metamorphic, and volcanic rocks, can be evaluated with the techniques described here, provided the list of minerals is small and their physical properties can be determined. In most carbonate reservoirs, the lithology is usually reasonably well known from sample descriptions or can be determined from log response. This is not true in sandstones because the mineral makeup of the sand is NOT usually described in much detail. There is a universal trend to give sandstones the physical properties of pure quartz, but this is almost universally NOT appropriate. Most sandstones contain other minerals such as mica, volcanic rock fragments, calcite, dolomite, anhydrite, and ferrous minerals, as well as the shale and clay described above. All of these minerals have different density, acoustic, and neutron properties than quartz. If a sandstone is assumed to be pure quartz when it is not, the commonly used properties of quartz will provide pessimistic porosity answers. Thus, authors and service company manuals that present mineral properties for “sandstone” are misleading their audience into believing these properties are constant. In more than 40 years of petrophysical analysis, I have never seen a thin section or XRD report that gave an assay of 100% quartz in any petroleum reservoir. A 100% quartz sand is very rare. If anyone doubts this statement, look at the PEF curve. If it reads more than 1.8, you have “quartz plus other things” in your sandstone. There is a story (it may even be true) that reserves for the early North Sea discoveries were seriously underestimated because the mica in the sands was not accounted for properly. The engineers used density log porosity without correcting for the real matrix density. If true, good engineering practice would have undersized all the offshore equipment and early cash flow and rate of return on investment would have been significantly reduced. If the myth that sandstone is pure quartz is perpetuated, there will be more economic blunders of this type. To further confuse the uninitiated, many logs show data on a "porosity" scale. These log curves are transforms of some measured physical property to an approximate porosity based on some arbitrary parameters. Examples are density, neutron, or sonic porosity on so-called Sandstone, Limestone, or Dolomite porosity scales. Porosity as defined by these transforms is only directly useful if there is no shale, the scale matches the rock mineralogy. and there are no accessory minerals. Real reservoirs are rarely this simple. DO NOT use these porosity transforms without further analysis unless all the arbitrary assumptions used to create them match exactly the rock you are analyzing. Some people call these porosity curves an “interpretation”. They are not. They are merely a transform of the raw data to a more attractive scale. The difference between a transform and an interpretation is critical. Interpretation infers some intelligent thought went into creating and understanding the result. The service company running the log does not provide interpretations. YOU are the interpreter. There are endless cases where a transform to an inappropriate porosity scale has caused millions in losses due to poorly informed analysts who see “gas cross over” when there is no gas, or who read porosity directly from the transform and either seriously over estimate or under estimate reservoir effective porosity. In spite of these comments, a number of charts and tables in this Chapter and elsewhere in this Handbook show the word "sandstone' when they really should say "quartz". I have not edited the charts and tables taken from common sources, such as service company chart books, so the common usage of incorrect terminology is repeated even here. It
should be noted also that this book uses the term "matrix
rock" to mean the solid, non-shale portion of a porous or
non-porous rock. In petrographic descriptions, "matrix"
is the clay between rock grains.
9.01
Visual Methods of Lithology Identification
FIGURE 9.01: Sandstone/Shale identification Visual identification of lithology is best explained by describing a few examples. In Figure 9.01, the density neutron log and the gamma ray are used to identify sandstone and shale as the basic rock components. This is evident because the logs are recorded on a sandstone scale, and the two logs read about the same porosity in the clean (non-shaly) zones. Therefore, sandstone is indicated. If the logs had been recorded on limestone scales, crossover of 6 to 7% porosity would have occurred - again indicating sandstone, or a gas filled limestone. If some other minerals were present, such as limestone, coal, or siderite beds, then the specific log characteristics for those zones would stand out from the average sand and shale values. For example, a limestone bed would be evident on the density log by its apparent low or negative porosity (high density) when recorded on a sandstone scale. Siderite, which is an iron rich mineral, may show a low or negative porosity as well. It would be hard to distinguish, by visual means alone, between it and a limestone stringer in a sandstone sequence.
A more complex example is shown in Figure 9.03, in which coal and dolomite appear as well as sands and shales. FIGURE 9.02: Lithology from a single log is usually impossible Figure 9.03 shows a carbonate sequence from Classic Example 2, logged on a limestone scale with a density neutron log. Here the density and neutron both read similar values in clean, non-shaly limestone. There is a separation of 8 to 12 porosity units if the carbonate is a pure dolomite and is fairly clean. Shale can also have this kind of separation, so the gamma ray log is necessary to define dolomite from shale. Anhydrite shows a separation of 12 to 20 porosity units.
On a limestone scale log, the rules for mineral identification are pretty simple: * PHIN near PHID, low GR = limestone or gas in dolomite Remember, the above rules are for compatible scale density neutron logs recorded on limestone units scale. Add 0.07 to all the separation values if logs are recorded on sandstone scale. A neat tool developed by Yalcin Pekiner for learning the separation rules is located HERE. When the photo-electric curve is available, the density neutron rules can be enhanced by the following when GR is low (not too shaly): *
PE near 2 = sandstone The most ambiguous case is radioactive dolomite (PHIN > PHID, PE near 3, high GR) as this is often mistaken for shale. The thorium curve on a natural gamma ray spectral log can be used to differentiate – dolomite has low thorium, shale has high thorium. Radioactive sandstone (feldspar sands, granite wash) has high GR but density neutron separation and PE follow the rules given above. The thorium curve will also help as feldspar sands have low thorium (and high potassium) while shale has high thorium (and may have high potassium as well). Radioactive limestones are usually fractured but otherwise they obey the normal PE and separation rules since the radioactivity is from uranium salts and not feldspar. Again the spectral gamma ray is helpful, showing high uranium. If a uranium corrected gamma ray curve (CGR) is available, compare it to the total gamma ray curve (SGR or GR). CGR < SGR means uranium is present. Crossplots in Section 9.09 may help to honour these rules for radioactive minerals.
FIGURES 9.04 and 9.05: Sonic and density-neutron logs in evaporite sequence The evaporite sequence shown in Figure 9.04 and 9.05 can be interpreted by comparing the absolute log values to the data in Section 9.16, which contains usual log readings for the commonly encountered sedimentary minerals. Lithology identification is aided materially in evaporites by sonic log data (Figure 9.04). The matrix rock values for common sedimentary minerals is given in Section 9.16 and for igneous and metamorphic rocks in Section 9.17. As with shaly sand sequences, the sonic is not especially helpful in carbonate sequences, as far as visual identification is concerned. Another complex lithology example is shown in Figure 9.06; layers are well defined by the density neutron separation.
The most recent advances in visual identification of lithology make use of the photoelectric effect (PE) curve of the lithodensity log, in combination with the density and neutron log readings. The rules were given earlier in this Section. The photoelectric effect for pure minerals are listed in Table 9.01, and vary slightly with porosity and hydrocarbon type. The amount of variation is shown in Figure 9.07. Even though literature suggests strongly that there is no porosity effect, and most formulae make this assumption, the variation in water filled high porosity can be as much as 15% of the PE value. This is not trivial and should be considered when using PE data for estimation of lithology in dual mineral situations. It is fortunate that the PE values for common minerals (quartz, calcite, and dolomite) do not overlap, even with porosity effect, so interpretation is relatively clear cut. However, anhydrite falls near calcite and must be distinguished by its characteristic high density. Shales and shaly sands can fall anywhere in the PE spectrum depending on the clay type and the amount of shale. A pure shale of illite would fall near dolomite, but a pure shale of kaolinite or montmorillonite would fall near sandstone. Chlorite shales have abnormally high PE values and can usually be distinguished. Thus shaly sands can fall anywhere between quartz (PE = 1.81) and chlorite (PE = 6.3) depending on clay type and shale content.
FIGURE 9.08: PE values and density neutron separation help define lithology Note that the density neutron log is in limestone units and no gas effect is present. If gas effect is present, then ambiguity still persists. If shale is present, the overlay should not be used. An example of real log data is provided in Figure 9.09. A similar example for shale identification is shown in Figure 9.10.
9.02
Lithology from Matrix Density WHERE: By rearranging the density response equation, we can derive apparent matrix density, based on the final effective porosity, shale volume, and the density log reading. Sxo is assumed to be 1.0. The matrix density thus derived can be compared to the data in Table 9.01 to find an approximate lithology. If PHIe is derived from a crossplot method (with or without the use of the density data), or from any one-log shale corrected method, the matrix density (DENSMA) can be back calculated as follows.
Calculate
density log value from porosity log reading. Rearrange
the response equation to solve for DENSMA. WHERE: COMMENTS: The matrix density will be too low in gas zones and in rough hole. No obvious correction can be made to overcome this problem. Do not use apparent matrix densities under these conditions. The standard density neutron crossplot can be used to determine matrix density, as shown in Figure 9.11.
To
produce a lithology code on computer listings, we bracket the
DENSma values as follows: Evaporites
require special handling in computer programs and the above codes
should be expanded to include: COMMENTS:
To
produce the fraction of each mineral present in a two-mineral
model, the computed matrix density is interpolated linearly between
the two end point densities. This is merely a rearrangement of
a response equation written in terms of matrix density, where
all terms go to zero except the matrix term. WHERE: COMMENTS:
PARAMETERS:
DENSSH
2.50 - 2.83 2500 – 2830
DENSMA
2.
If we use the Vsh from the density neutron crossplot (Vsh = 0.59) This is an impossible result which suggests that the shale volume or the effective porosity is wrong. Calculated matrix density can thus be used as a quality control indicator when it exceeds reasonable bounds. 3.
Assume the two mineral model consists of quartz and dolomite and
the computed matrix density is 2680 Kg/m3. Then: Note that V1 + V2 + Vsh + PHIe = 1.0 and that V1 and V2 do not represent the fraction of each matrix rock relative to each other, but to the total bulk rock volume.
9.03
Lithology from Matrix Travel Time
1:
IF Vsh + PHIe < 0.95 WHERE: COMMENTS: The matrix travel time can be obtained graphically from the chart in Figure 9.12.
Lithology codes are more difficult to generate with sonic data then with density data.
COMMENTS: Should not be used in shallow unconsolidated sandstones.
Linear
interpolation of the DELTma value between the two end points of
a two mineral model can be accomplished in a fashion similar to
the matrix density described earlier. WHERE: COMMENTS: This solution guarantees that all components sum tp 1.00. To obtain V1 and V2 as comparative volumes, set Vrock = 1 in equations 2 and 3.
PARAMETERS:
DELTMA NUMERICAL
EXAMPLE: This value falls in the impossible area and is too high because the sonic log reads high compared to effective porosity found from the density neutron crossplot. If porosity was 0.16, the matrix travel time would be 183 usec/ft (close to the sandstone value). This is another quality control indicator, and in this example demonstrates a lack of coherence between the sonic and density neutron data, when the matrix, shale and fluid assumption are as given above. Either these parameters, or the log data, or the whole rock model are in error. 9.04
Secondary Porosity
Find
pseudo matrix travel time. These formulae define a pseudo sonic matrix travel time used to calculate sonic porosity with the same matrix as the density neutron porosity. Calculate
sonic porosity. Compare
to density neutron crossplot porosity. WHERE: COMMENTS: RECOMMENDED
PARAMETERS: NUMERICAL
EXAMPLE: Since PHIs2 > PHIe, there is no secondary porosity and PHIsec = 0.0. 2.
Assume DENSma2 = 2.79 gm/cc and DELT = 65 usec/ft 9.05
Lithology from Mlith-Nlith Method
1:
PHIdc = PHID - Vsh * PHIDSH WHERE: COMMENTS: These two variables are usually called M and N, but they can be confused with the cementation exponent M and the saturation exponent N, so we have changed their names to reduce confusion.
PARAMETERS:
Clean Quartz –
0.028 2650 182 0.802
0.623 1.82 4.8 The end points for the common minerals listed above are plotted in Figure 9.14.
If
the usual lithology is made up of two minerals, then the Mlith
and Nlith values can each be linearly interpolated to find the
fraction of the minerals. WHERE: COMMENTS: This solution guarantees that all components sum to 1.00. To obtain V1 and V2 as comparative volumes, set Vrock = 1 in equations 2 and 3.
If
the usual lithology is made up of two minerals, then the Mlith
and Nlith values can each be linearly interpolated to find the
fraction of the minerals. WHERE: COMMENTS:
RECOMMENDED
PARAMETERS:
If
the usual lithology is made up of three minerals, then the Mlith
and Nlith values can be linearly triangulated to find the fraction
of the minerals. WHERE: COMMENTS:
An alternate version of this model can be made by replacing Nlith with Plith = PE / (DENS - DENSW) - density in gm/cc. This avoids the use of the neutron log in cases where it has little lithology discrimination, such as in igneous rocks. RECOMMENDED
PARAMETERS: NUMERICAL
EXAMPLE: The closest values in the table represent dolomite (Mlith = 0.778 and Nlith = 0.516), so this interval is very likely dolomite.
9.06
Lithology from Alith-Klith Method The input data to these algorithms must be shale corrected, must be in limestone porosity units and must be in English units before processing begins.
1:
PHIdc = PHID - Vsh * PHIDSH |
When
a single porosity log is all that is available, it is difficult
and sometimes impossible to find a baseline that will indicate
lithology. Use a sample description or regional knowledge to define
rock matrix, then proceed to estimate porosity as described earlier.
On
a sonic log in a sand-shale sequence, there may be a few tight
sandstones which will record values close to the matrix value
of sandstone, which is 55.5 microseconds per foot (or 182 microseconds
per meter). Such tight sandstones may not be present in a well
and therefore it would be difficult to prove that the sequence
is sand-shale just from the sonic log alone. Data from Figure
9.02, from Classic Example 1, does not show any baselines suitable
for identifying lithology.
FIGURE
9.03: Lithology assessment in a carbonate sequence
Figure
9.06: Lithology analysis on density neutron log
The
visual overlay method for the neutron density and photoelectric
data works quite well for some two mineral situations. Again,
it is the separation between the curves that counts. A schematic
diagram, based on the data in Section 9.16, is shown in Figure
9.08.
