Petrophysics MSC Course Notes.
Petrophysics MSC Course Notes.
1.1 Introduction
The search or economic accumulations of oil and gas starts with the recognition of likely geological
provinces, progresses to seismic surveying, and the drilling of one or more wild-cat wells. If one is
lucky, these wells may encounter oil, and if that is the case, measurements made down the hole with
wireline or LWD (Logging While Drilling) tools are used to assess whether sufficient oil is present,
and whether it can be produced. Clearly, the evaluation of sub-surface formations requires the
combined efforts of geologists, petrophysicists, drilling engineers, reservoir engineers and
geophysicists.
The geologist is interested in the lithology, stratigraphy and depositional environment of the sub-
surface strata penetrated by the drilling bit. The exploration geologist uses formation evaluation data in a
number of wells to create a large scale image of the sub-surface geology by correlating responses
that are characteristic of a given formation or horizon between formations. This picture is very useful
when carrying out initial reservoir modelling and in the decision where to drill new wells. Later the
production geologist carries out much the same process with much more well information, and adds
any extra information that has been gathered to produce a detailed geological model of the reservoir and
related sub-surface formations. This model will be the basis of reservoir modelling, and all major
reservoir management decisions from primary drainage through to enhanced oil recovery and shut-down.
The petrophysicist’s job is to use all available information to analyze the physical and chemical
properties of the rocks in the sub-surface, and their component minerals, with particular emphasis
given to the amount and distribution of water, oil, and gas. The petrophysicist will use extensively
wireline and LWD log data and data from experiments done on cores extracted from the well, and will
occasionally use other sources of information such as engineering and production logs, as well as mud
logging data. Initially, it is the aim of the petrophysicist to differentiate between oil, gas and water
bearing formations, estimate the porosity of the formations and the approximate amount of
hydrocarbons present in each formation. Ultimately, the petrophysicist also uses laboratory data to
estimate how easy it will be to extract the hydrocarbons in place and, with the reservoir engineer, to
design reservoir management strategies to optimize long term oil recovery.
There is a large database of information available to both the geologist and the petrophysicist, and as
time passes the amount and variety of information increases. Table 1.1 summarizes a few of the main
measurement that a geologist or petrophysicist will have access to, arranged in approximate
chronological order. It is the responsibility of the wellsite geologist or engineer to ensure that all this
data is properly collected and recorded.
This course is designed to introduce you to wireline and LWD formation evaluation logs and the simple
analysis of the data that they provide. The following sections cover the theory behind the operation of
each tool, how each tool operates, its advantages and limitations, and its main applications. There are a
huge number of tools available in the industry, too many to be covered individually. However, all are
variations on a small number of generic tool types, which will be described.
It should be remembered at all times that the main job of the petrophysicist is to evaluate the amount
of hydrocarbons in place in the reservoir. Hence, the evaluation sequence for a straightforward
reservoir will be as follows:
2. For the reservoir intervals only, distinguish between hydrocarbons and water filling the
pores, hence calculate water saturation in reservoir rocks
(Hydrocarbons are electrical insulators, while water conducts.)
3. For the hydrocarbon fraction, distinguish between oil and gas, hence calculate gas and
oil saturations in reservoir rocks
(Gas has a much lower density than oil.)
These processes can take extremely long periods of time. Most formations that contain reservoirs are
sedimentary rocks, where the deposition of organically rich material has been followed by clean
sandstones that form high porosity well connected pore systems, and are subsequently capped by
shales with very low permeabilities. Here the burial of the initial deposition provides the pressures and
temperatures to produce hydrocarbons. The hydrocarbons are less dense than water, so migrate
upwards into the sandstones, replacing the water that originally occupied the reservoir sandstone,
where the hydrocarbons are constrained from rising further by the shale cap. The hydrocarbon then
waits until we decide to produce it.
The depositional and post-depositional history of the reservoir rock, and particularly its diagenetic
history (compaction, cementation and dissolution), all contribute to the mineralogical composition of
the rock, and hence its grain size distribution, porosity, pore size distribution and the connectivity of
its pores. Note that in the process of migration the hydrocarbon replaces water in the reservoir rock
because it is less dense than water. In practice, the replacement is almost never complete, with some
water associated with even the best oil accumulations. The reason for the remaining water is that the
grains comprising the reservoir rock are usually water-wet, i.e., having a chemical preference to be
covered in water rather than hydrocarbon, hence they retain a thin film of water when the hydrocarbon
replaces most of the water in the pores. Oil-wet rocks do exist, and the ability to distinguish between
oil and water wet rocks is extremely important in reservoir management, especially in the final stages
of reservoir production. In general, any given reservoir rock the pore space will be occupied by a
water saturation Sw, a gas saturation Sg, and an oil saturation So.
Since gas is less dense than oil, which is less dense than water, the fluids separate in hydrocarbon
reservoirs with the gas occurring just below the trapping lithology, oil a little deeper, and water at the
bottom. The fluids are commonly immiscible and so we can define a gas-oil contact (GOC) and an oil-
water contact (OWC). Since, gravity is the force that separates the fluids into these layers, the GOC
and OWC are horizontal providing that horizontal and vertical permeability is good in the reservoir
and there are no complicating structures or fractures. Note that it is not compulsory to have all three
fluids occurring together. Hence in gas reservoirs the oil is missing and there is a gas-water contact
(GWC). Similarly, oil reservoirs can exist without a gas cap.
The calculation of hydrocarbon volume requires us to know the volume of the formations containing
the hydrocarbons, the porosity of each formation, and the hydrocarbon saturation in each formation. In
practice each reservoir will be made up of a number of zones each with its own thickness, areal extent,
porosity and hydrocarbon saturation. For example, reservoir sandstones may alternate with non-
reservoir shales, such that each zone is partitioned. Such zonation is mainly controlled by lithology.
Hence, it is an early requirement to identify the lithologies in a particular well, identify which
formations have the required porosity to enable it to be a reservoir rock, and determine whether the
formation contains hydrocarbons. Reservoir rocks containing hydrocarbons are allocated a zone code.
The volume of reservoir rock in a single zone depends upon the area of the zone A, and the thickness
of reservoir rock in the zone h. The area is obtained usually from seismic data (from the reservoir
geologist), and is the only data used in the calculation of hydrocarbon volumes in place that is not
derived from petrophysical techniques. The thickness of reservoir rock is derived from the zonation of
the reservoir based upon an initial lithological interpretation and zonation of the reservoir from the logs.
The bulk volume of the reservoir Vbulk =A × h.
The majority of this volume is occupied by the solid rock matrix, and the remainder is made up of the
pore space between the minerals. The relative amount of pore space to the bulk volume is denoted by
the porosity, where the porosity is the fraction of the bulk volume occupied by pore volume, and is
expressed as a fraction or as a percentage; Φ =Vpore/Vbulk. However, note that the fractional form is
used in ALL calculations. The pore volume in any given zone is therefore Vpore= Φ x A x h.
In general the porosity is completely occupied by either water or hydrocarbon, where the saturation of
the water is Sw, and that of the hydrocarbon is Sh, and Sw + Sh = 1. In most reservoirs the
hydrocarbon has replaced all the water that it is possible to replace, and under these conditions the
water saturation is termed the irreducible water saturation Swi. Now we can write the hydrocarbon
saturation as Sh = (1 – Sw). Hence the volume of hydrocarbons in place can be written as
The determination of this value is the primary job of the petrophysicist, and requires a lithological
assessment and zonation of the reservoir. At a later stage the petrophysicist may also be called upon to
assess the permeability of the reservoir under various conditions. However, the primary function of the
petrophysicist is to assess the amount of hydrocarbons initially in place.
All the parameters in Eq. (1.1) except the area are derived from measurements made in the borehole
using wireline tools or increasingly using data obtained from tools that measure the rock formations
during drilling (measurement/logging while drilling: MWD/LWD).
A
(1-Sw)
Hydrocarbon Volume Φ
Ah Φ (1-Sw) Sw h
Water Volume Ah Φ S w
Ah (1- Φ)
Matrix Volume
Equation (1.1) gives the volume of hydrocarbons in ft3 or m3, and the calculated values are often
unwieldy. The oil industry uses a range of industry standard units, in which all calculations should be
carried out. Commonly the volume of oil is measured in barrels, the amount of gas in cubic feet,
thickness in feet, and area in acres. Conversions for these are given in Table 1.2 below.
Hence, for an oil reservoir zone of A acres and h feet thickness, the volume of oil in place (OIP) is
and for a gas reservoir of the same dimensions, the volume of gas in place (GIP) is
Note that sometimes these values are phrased as the amount of oil originally in place (OOIP) and gas
originally in place (GOIP).
If the area is not known, the amounts of oil or gas are quoted in bbl/acre or cu.ft/acre respectively,
and if the reservoir zone thickness is also not known, the values are given in bbl/acre.foot and cu.ft/acre.ft
respectively.
These calculations are done for reservoir zones, which are thicknesses of formation of the same
lithology, with similar cleanness (i.e., containing little clay), porosity, permeability, and hydrocarbon
saturation (i.e., deep resistivity). The zonation of a reservoir is a qualitative exercise that often depends
upon the individual petrophysicist. There will often be a minimum thickness for a zone that depends
upon the study, however zones are rarely defined less than 5 ft. thick.
The thickness of the reservoir zone used in this calculation is the net sand. The net sand is the
thickness of clean, permeable, hydrocarbon-containing sand in the reservoir zone. The gross is the
thickness of the reservoir zone irrespective of whether it is clean, its permeability or its fluid
saturation. The net to gross ratio (thickness of net sand divided by the gross thickness) is often used to
represent the quality of a reservoir zone.
Also note that the oil and gas will be at raised temperature and pressure in the reservoir. The
compressibility of oil and especially gas, and their coefficients of expansion with temperature means
that they will occupy different volumes at surface pressure and temperature conditions, or those
present in the stock tank at the surface. For this reason reserves are often quoted corrected for the
changes in temperature and pressure at the conditions of the stock tank. If this has been done the stock
tank oil and gas originally in place is given as STOOIP and STGOIP.
The expansion or reduction in volume undergone by oil and gas as its temperature and pressure
conditions change from that in the reservoir to that in the stock tank depend upon the changes in
pressure and temperature and the composition of the oil or gas. The change is expressed by what are
called formation volume factors. The oil formation volume factor Bo is the ratio of the volume of a
standard mass of oil at reservoir conditions to that at stock tank conditions, and has no units. Hence,
we can calculate now the amount of oil originally in place in the reservoir when measured at the
pressure and temperature conditions prevailing in the stock tank (i.e., during production).
Similarly, the gas formation volume factor Bg is the ratio of the volume of a standard mass of gas at
reservoir conditions to that at stock tank conditions, and also has no units. Hence, we can calculate
now the amount of gas originally in place in the reservoir when measured at the pressure and
temperature conditions prevailing in the stock tank (i.e., during production).
Note that Bo>1, hence the volume of oil is less at the surface than at depth. This is because the
compressibility of oil with pressure for the range of pressures usually found in reservoirs is less than
the thermal expansion of the oil caused by reservoir temperatures. Therefore, when hot, pressurized oil at
depth is brought to the cooler, lower pressure stock tank it contracts due to the lowering of
temperature more than it expands due to the lowering of pressure.
Whereas, the calculation for reservoir OOIP from Eq. (1.2) would have given 244.377 Mbbl.
Note that Bg<<1, hence the volume of gas is much greater at the surface than at depth. This is because
the compressibility of gas with pressure for the range of pressures usually found in reservoirs is enormous
compared to the thermal expansion of the gas caused by reservoir temperatures. Therefore, when hot,
pressurized gas at depth is brought to the cooler, lower pressure stock tank it expands considerably
more due to the lowering of pressure more than it contracts due to the lowering of temperature. This is
the opposite of the case for oil.
Whereas, the calculation for GOIP from Eq. (1.3) would have given 1.372 billion cu.ft. The huge
difference in the volume of the gas at stock tank conditions compared with that at depth indicates the
importance of taking full account of the formation volume factor for gas.
It should, however, be noted that gas and oil cannot usually be taken as independent as in the two
examples above. In practice, gas is dissolved in oil, and comes out of solution as the oil is brought to
the surface. This results in additional shrinkage in the apparent oil volume, and this is taken into
account in quoted figures for the oil formation volume factor. The reservoir then also produces gas,
and the exsolved gas adds to the volume of the gas produced.
The complex interplay between gas and oil phase behaviours will be covered in formation evaluation
course later in the MSc.
1.4 Uncertainties
Note that errors in any of the measurements leads to errors in the final value, and that the porosity and
water saturation are in some ways inter-dependent, which can lead to larger error than might otherwise
be expected.
The error in the area of a reservoir zone is usually quite small because seismic surveys can delineate
the extent of reservoirs with great accuracy now. A greater error can be introduced from petrophysical
measurements. These values are derived from a small number of wells that intersect the reservoir.
There arises, therefore, the possibility that the well information is not representative of the reservoir as
a whole. Clearly, the greater the number of wells, the smaller the possible errors from this source.
However, early in the development of a field, when the result of reservoir management decisions can
have a large impact upon the reserves and life of a field, the amount of petrophysical data is small, and
the resulting errors can be quite large. It is therefore, always important to realize the impact that
uncertainty in the knowledge of the petrophysical parameters can have in the final value of STOOIP.
Table 1.3 shows the impact that a 10% variation in the petrophysical parameters can have in the
calculation of STOOIP.
These errors can be analyzed statistically to give a probability curve from which the most likely value
of STOOIP can be judged. As more and better quality data become available, we can reduce the
possible error on the STOOIP analysis, generating probability curves with less variation. Figure 1.2
shows an example of probability curves for an initial appraisal, followed by two other curves made 8
and 15 years later. Note that there has been great improvement in the first 8 years, when appraisal
drilling added much more data to the data base, followed by less improvement in the subsequent 7
years, when appraisal drilling was reduced significantly, and hence there was less new data added to
the database. As with all things, there is an economic balance between added value generated from
new data and the cost of obtaining the new data.
Figure 1.2 is read by taking the ordinate (y-axis) and reading across. For example in 1972 there was
100% probability that the reservoir contained greater than 80 Mbbl, and 75% probability that it
contained more than 250 Mbbl. By 1980 there was 100% probability of the reservoir containing
greater than 220 Mbbl and 75% probability of it containing greater than 310 Mbbl of oil. Thus, as time
progresses the range of possible STOOIP decreases from a possible range of 80 Mbbl<STOOIP<600
Mbbl in 1972 to 260 Mbbl <STOOIP<470 Mbbl in 1987.
The data represented by Fig. 1.2 is shown in the Table 1.3 below.
Most critical financial and reservoir management decisions usually have to be made early in the life of a
field when data is at a premium and errors in the estimated STOOIP are greatest. This is particularly true
of offshore fields, where the cost of top-side production structures is a significant factor in the
economic viability of the field. For example, globally, the Norway and UK sectors of the North Sea
are the two most expensive oil provinces to produce from due to their deep off-shore nature combined
with inhospitable conditions. In these cases it is critical to obtain as much high quality data as possible
from the first few appraisal wells in each prospective reservoir.
Since it is clear that the main interests of petrophysicists are porosity and permeability, the following
few chapters (Chapters 2 and 3) will deal with these parameters. Fluid saturations are also critical, and
these depend upon capillary pressure, which is described in Chapter 4.
2. POROSITY
2.1 Theory
The porosity of a rock is the fraction of the volume of space between the solid particles of the rock to
the total rock volume. The space includes all pores, cracks, vugs, inter- and intra-crystalline spaces.
The porosity is conventionally given the symbol ф, and is expressed either as a fraction varying
between 0 and 1, or a percentage varying between 0% and 100%. Sometimes porosity is expressed in
‘porosity units’, which are the same as percent (i.e., 100 porosity units (pu) = 100%).
However, the fractional form is ALWAYS used in calculations.
Porosity is calculated using the relationship
It should be noted that the porosity does not give any information concerning pore sizes, their
distribution, and their degree of connectivity. Thus, rocks of the same porosity can have widely
different physical properties. An example of this might be a carbonate rock and a sandstone. Each
could have a porosity of 0.2, but carbonate pores are often very unconnected resulting in its
permeability being much lower than that of the sandstone. A range of differently defined porosities are
recognized and used within the hydrocarbon industry. For rocks these are:
2.2
Controls on Porosity
The initial (pre-diagenesis) porosity is affected by four major microstructural parameters. These are
grain size, grain packing, particle shape, and the distribution of grain sizes. However, the initial
porosity is rarely that found in real rocks, as these have subsequently been affected by secondary
controls on porosity such as compaction and geochemical diagenetic processes. This section briefly
reviews these controls.
The calculations of these ideal porosities is relatively simple. For example, taking the cubic
arrangement of identical spheres of radius r occupying a cubic unit cell of length L, as shown in Fig.
2.2, the following calculation is possible.
L r
The bulk volume of the cell Vbulk = L3, and the number of spheres in the cell n =(L/2r)3. Hence the
volume of the matrix V matrix = (4 n r3 )/3 = (L/2r)3 (4 r3)/3 = ( L3)/6. The porosity can now be
calculated from Eq. (2.1) as
(2.2)
which is independent of the sphere size. Most of the other values in Table 2.1 can be calculated
similarly, although with a little more difficulty as a result of their different packing geometries. There
are 6 different ways that identical spheres can be packed, and these are shown in Fig. 2.1.
The equilibrium porosity of a porous material composed of a random packing of spherical grains is
dependent upon the stability given to the rock by frictional and cohesive forces operating between
individual grains. These forces are proportional to the exposed surface area of the grains. The specific
surface area (exposed grain surface area per unit solid volume) is inversely proportional to grain size.
This indicates that, when all other factors are equal, a given weight of coarse grains will be stabilized
at a lower porosity than the same weight of finer grains. For a sedimentary rock composed of a given
single grain size this general rule is borne out in Figure 2.3. It can be seen that the increase in porosity
only becomes significant at grain sizes lower than 100 µm, and for some recent sediments porosities
up to 0.8 have been measured. As grain size increases past 100 µm, the frictional forces decrease and
the porosity decreases until a limit is reached that represents random frictionless packing, which
occurs at 0.399 porosity, and is independent of grain size. No further loss of porosity is possible for
randomly packed spheres, unless the grains undergo irreversible deformation due to dissolution-
recrystallisation, fracture, or plastic flow, and all such decreases in porosity are termed compaction.
0.8
Angular Grains
Rounded Grains
0.7
0.6
0.5
0.4
Limit of uncompacted
random packing
0.3
0 100 200 300 400 500
Grain Diameter (microns)
Figure 2.3 Relation between porosity, grain size and grain shape
Rarely, in borehole petrophysics do we need to look at accurate grain size determinations as many of
the tools that we use have minimum vertical resolutions of the order of tens of centimetres. However,
as well logs are correlated to core logging it is well to bear in mind the agreed semi-quantitative
classifications for grain size in siliclastic and carbonate rocks (Tables 2.2 and 2.3).
Table 2.2 Siliclastic grain size definitions Table 2.3 Carbonate grain size definitions
Category Median Grain Size (microns)
Category Median Grain Size (mi)
L
Gravel
400
2000
M
Very Coarse
200
1000
F
Coarse
100
500
VF
Medium
50
250
EF
Fine
125 Symbol Pore Size (microns)
Very Fine B <100
C 100-200
62
D >200
Silt
Matrix Code Texture
I Compact
II Chalky
III Sucrose
grains are removed and replaced with a single solid larger grain, the porosity will be reduced because
both the small grains and their associated porosity have been replaced with solid material. The solid
lines GR and RF or RM in Figure 2.4 represent the theoretical curves for both processes. Note that as
the disparity between the grain sizes increases from 6:3 to 50:5 the actual porosity approaches the
theoretical lines. Note also that the position of the minimum porosity is not sensitive to the grain
diameter ratio. This minimum occurs at approximately 20 to 30% of the smaller particle diameter. In
real rocks we have a continuous spectrum of grain sizes, and these can give rise to a complex scenario,
where fractal concepts become useful.
(a) (b)
2 2
F
G M G
1 R 1 R
Figure 2.4 Apparent volume for assemblages of spheres of different diameter ratios.
In practice such detail is rarely entered into. Grain size distributions can be inferred from mercury
porisimetry measurements that will be studied later in the MSc course, or by direct sieving of a
disaggregated sample. In the sieving technique, the dried, weighed and disaggregated sample is placed
in the topmost of a stack of sieves with decreasing mesh size D. A motor vibrates the stack such that
given particle sizes remain in each of the sieves. The cumulative percentage weight of these remains
are plotted against the mesh size of the sieves D, resulting in a distribution curve similar to that in Fig.
2.5. The degree of sorting (grain size distribution) is expressed by the Trask Coefficient so, which is
expressed as
(2.3)
where: D25 = mesh size (grain diameter) at 25% of sample weight (larger grain fraction).
D75 = mesh size (grain diameter) at 75% of sample weight (all but smallest grain fraction).
100
75
50
25
D25 D75
0
0
Mesh Diameter Increasing
Figure 2.5 Cumulative grain size distributions by sieved weight.
In general the porosity of a sandstone decreases as the degree of sorting decreases, where the degree of
sorting is expressed as a Trask coefficient or classified according to Table 2.5.
There are at least 4 common methods of measuring the porosity of a core. These are:
• Buoyancy
• Helium porisimetry
• Fluid saturation
• Mercury porosimetry
The first three will be described here, and the last will be described in Chapter 4.
(2.4)
Note that the mean grain density of the sample can also be calculated from this data using the
relationship
(2.5)
This method is somewhat similar to the saturation method, and also measures the effective (connected)
porosity.
• Clean and dry the rock.
• Weigh the rock in its dry state to give the dry weight, Wdry.
• Fully saturate the rock in a wetting fluid as before.
• Weigh the saturated sample suspended in a bath of the same fluid with which it was saturated to
give its suspended weight, Wsus. This is shown in Fig. 2.6. Note that a cradle is needed to
suspend the sample in the fluid, and this has a weight Wcradle when in the fluid. So the actual
measured weight when the sample and cradle are suspended in the fluid is (Wsus + Wcradle). The effect
of the cradle must be taken away, hence the weight of the cradle when suspended in the fluid (up to
the same level), Wcradle, must be taken account of in the final calculation
• Determine the bulk volume of the rock, Vbulk as before.
• Determine the density ρfluid of the saturating fluid by weighing a known volume of it.
Weighing Device
Cradle with
Level Mark
Core
Sample
(2.6)
(2.7)
This method uses the apparatus shown in Fig. 2.7, together with the following procedure.
Pressure
Gauge
To
Sample Atmosphere
Gas
Supply
Valve 1 Valve 2 Valve 3
Chamber A Chamber B
The drop of pressure depends upon how much space there is in Chamber B, and that depends on how
much of Chamber B is occupied by solid rock particles. The measured gas pressures Pi, Patm and Pf
alone are not sufficient to obtain the volume of the sample rock matrix. However, we can calibrate the
system by putting a range of metal cylinders of accurately known volume into Chamber B and
repeating the experiment. When this is done, the calibration constant, and the pressures allow the
volume of the rock matrix Vmatrix to be obtained. The porosity is then calculated using
(2.8)
3. PERMEABILITY
3.1 Theory
The permeability of a rock is a measure of the ease with which the rock will permit the passage of
fluids.
The fundamental physical law which governs this is called the Navier-Stokes equation, which is very
complex. For the purposes of flow in rocks we can usually assume that the flow is laminar, and this
assumption allows great simplification in the equations.
It should also be noted that the permeability to a single fluid is different to the permeability where
more than one fluid phase is flowing. When there are two or more immiscible fluid phases flowing we
use relative permeability, which will be introduced in this section, but covered in much more detail on
the Formation Evaluation course later in the MSc.
The fluid flow through a cylindrical tube is expressed by Poiseuille’s equation, which is a
simplification of Navier-Stokes equation for the particular geometry, laminar flow, and
uncompressible fluids. This equation can be written as
(3.1)
About 150 years ago Darcy carried out simple experiments on packs of sand, and hence developed an
empirical formula that remains the main permeability formula in use in the oil industry today. Darcy
used the apparatus shown in Fig. 3.1, where he used a vertical sand pack through which water flowed
under the influence of gravity while measuring the fluid pressures at the top and bottom of the pack by
the heights of manometers. Here the difference in fluid pressures can be calculated from h1–h2
providing the density of the fluid is known. It has since been validated for most rock types and certain
common fluids. Darcy’s formula can be expressed as
(3.2)
Note, either the first set of quoted units can be used (c.g.s system) as in the oil industry, or the second
set (S.I. units) as in academic research, but one must use the units consistently.
The units of permeability are the darcy, D, and m2, where 1 D = 0.9869× 10-12 m2. One darcy is the
permeability of a sample 1 cm long with a cross-sectional area of 1 cm2, when a pressure difference of
1 dyne/cm2 between the ends of the sample causes a fluid with a dynamic viscosity of 1 poise to flow
at a rate of 1 cm3/s (Fig. 3.2). In geological applications the darcy is commonly too large for practical
purposes, so the millidarcy (mD) is used, where 1000 mD = 1D.
Fl
Pi Po
Flow Q Flow Q
A
uid Pressures L
(3.3)
(3.4)
where: ∆V = the volume of gas flowed in time ∆T measured at atmospheric pressure (cm3 )
∆T = the time period over which flow is measured (s)
However, we can make some generalizations if all other factors are held constant:
The permeability of rocks varies enormously, from 1 nanodarcy, nD (1× 10-9 D) to 1 microdarcy, µD
(1× 10-6 D) for granites, shales and clays that form cap-rocks or compartmentalize a reservoir, to
several darcies for extremely good reservoir rocks. In general a cut-off of 1 mD is applied to reservoir
rocks, below which the rock is not considered as a reservoir rock unless unusual circumstances apply
(e.g., it is a fractured reservoir). For reservoir rocks permeabilities can be classified as in Table 3.1
below.
<10 Fair
10 – 100 High
100 – 1000 Very High
>1000 Exceptional
At this point we must make a distinction between the use of gaseous fluids and the use of liquids. In
the case of liquids the measurement is relatively straightforward as the requirement for laminar flow
and incompressibility of the fluid are almost always met at surface geological conditions. If one wants to
use gas as the fluid, as is commonly done in the industry, there are two complications:
Gas is a compressible fluid, hence if gas is flowing at the same mass per unit time through the core, it
will actually be travelling more slowly when measured in volumes per time at the input (high pressure)
end of the sample because it is compressed into a smaller volume, than at the output end (low
pressure) where it expands. The equation used to calculate the permeability value from the measured
parameters has to be modified to take the gas compression into account.
At low gas pressures, there can be very few molecules of gas occupying some of the smaller pores. If
this happens, the laws that we are using breakdown, and their use causes an overestimation in the
permeability. This is known as gas slippage or the Klinkenberg Effect. The problem becomes smaller
as the pressure is increased because the gas is compressed and there are more gas molecules per unit
volume, and does not arise in liquids because liquids are very much denser than gasses.
Gas slippage is corrected for by making permeability measurements with gas at multiple pressure
differences and constructing a graph of the measured apparent permeability against the reciprocal of
the mean pressure in the core. If the input gas pressure is Pi and the output pressure is Po, then the
permeability is plotted as a function of 1/Pav = 2/(Pi + Po), as in Fig. 3.3. The points should now lie on a
straight line, which intersects the y-axis at 1/Pav = 0. This value is called the Klinkenberg
permeability, and effectively represents the permeability at which the gas (which is near to a perfect
gas) is compressed by infinite pressure and becomes a near perfect liquid. It is because of this that the
klinkenberg permeability is often given the symbol kL.
The klinkenberg permeability is very commonly used within the oil industry, and should approximate
very well to the permeability of the sample measured with liquid flowing through it. It should be noted
that the correction cannot be ignored, especially in tight rocks, as it can lead to corrections of up to
100%. In general, the correction is smaller for higher permeability rocks containing larger pores.
100
80
60
40
Klinkenberg
Permeability, k L
20
0
0 10 20 30 40 50
Reciprocal Mean Pressure, 1/Pav, psi
Permeability is generally anisotropic in a rock, partly because of depositional effects, and partly
because of the in-situ stress field in the crust. To account for this, permeability measurements are
made both parallel to and perpendicular with bedding. The permeability perpendicular to bedding will
be about a third to half of that parallel to bedding. Clearly this has implications for extracting oil from a
reservoir, as oil would usually much rather travel laterally than vertically.
If the rock contains one fluid, the rock permeability is maximum, and this value is called the absolute
permeability.
If there are two fluids present, the permeabilities of each fluid depend upon the saturation of each
fluid, and can be plotted against the saturation of the fluid, as in Fig. 3.4. These are called effective
permeabilities. Both effective permeabilities are always less than the absolute permeability of the rock
and their sum is also always less than the absolute permeability of the rock. The individual effective
permeabilities are most often expressed as a fraction of the absolute permeability of the rock to either
of the two fluids when present at 100% saturation, and these are called relative permeabilities.
Hence, if 100% water occupies the rock, the absolute permeability to water is kaw, and the same applies
for 100% saturations with oil (kao), and gas (kag). If any two, or all three, of these fluids are present
together in the rock at some partial saturation Sw, So and Sg, we can measure their effective
permeabilities, which are k ew, k eo, and k eg, which will all be less than their absolute values. We
can define and calculate the relative permeability values by expressing the effective permeabilities as a
fraction of some base permeability, which is arbitrary but usually the absolute permeability of one of
the fluids present. For example, if we take kao as the base permeability, the relative permeabilities are:
Note that the precise value of the relative permeabilities depends upon the base permeability with
which they are calculated, and this should always be quoted whenever relative permeabilities are used.
Referring to Fig. 3.4, it can be seen that the effective and hence relative permeability of a given fluid
decreases as the saturation of that fluid decreases, and that there is a threshold value of saturation of
any given fluid that needs to be present before that fluid will move. This last point is on the one hand
intuitive, as one would expect the need for sufficient of a given fluid to be present before a connected,
flowable pathway could come into being, and on the other hand critical, because it implies that fluids
become trapped (unmovable) in a rock when they are still present in significant amounts. In the figure,
oil is immobile until its saturation is about 20%, indicating that we cannot produce from zones that
contain less than 20% oil, and that we will not be able to produce the last 20% of oil from zones which
initially have higher oil saturations. This is known as the residual oil saturation, Sor.
1.0 1.0
0.8 0.8
kro
0.6 0.6
Swi Sor
0.4 0.4
k rw
0.2 0.2
0 0
0 0.25 0.5 0.75 1.0
Water Saturation
Figure 3.4 Relative permeability curves for an oil/water system that is water wet.
The same applied to the water, whose immobile fraction is termed the irreducible water saturation,
Swi, and gas, whose immobile fraction is termed the trapped gas saturation, Sgt.
There is a point on the plot where the curve intersect. Here the permeability to each fluid is the same,
and both fluids are equally easily produced. As the oil saturation increases, the permeability to oil
increases, and that to water decreases, and vice versa. Hence it is apparent that in oil reservoirs it is
important to avoid the production of water as not only does it not make money, but an increasing
water cut reduces the permeability of the reservoir to the oil, making oil more difficult to produce.
For the arrangement shown in Fig. 3.5, we can write the flow equation
(3.5)
rw
Where, the internal area of the borehole has been
substituted for the area A. It can be seen that for any instant
in time, when the formation pressure P f and the well
pressure Pw are constant, and for constant fluid viscosity, µ,
the productivity of the well will be proportional to the
permeability and the interval of production, h, with C being
the coefficient of proportionality.
Pf Pw h
Figure 3.5 Well productivity.
The most obvious control on permeability is porosity. This is because larger porosities mean that there
are many more and broader pathways for fluid flow. Almost invariably, a plot of permeability (on a
logarithmic scale) against porosity for a formation results in a clear trend with a degree of scatter
associated with the other influences controlling the permeability. For the best results these poroperm
cross-plots should be constructed for clearly defined lithologies or reservoir zones. If a cross-plot is
constructed for a whole well with widely varying lithologies, the result is often a disappointing cloud
of data in which the individual trends are not apparent. Figure 3.6 shows a poroperm cross-plot for a
clean sandstone and a carbonate.
100 100
10 10
1 1
0 10 20 30 40 0 10 20 30 40
Porosity, % Porosity, %
It is clear from this figure that the permeability of the sandstone is extremely well controlled by the
porosity (although usually there is more scatter than in this figure), whereas the carbonate has a more
diffuse cloud indicating that porosity has an influence, but there are other major factors controlling the
permeability. In the case of carbonates (and some volcanic rocks such as pumice), there can exist high
porosities that do not give rise to high permeabilities because the connectivity of the vugs that make up
the pore spaces are poorly connected.
Poroperm trends for different lithologies can be plotted together, and form a map of poroperm
relationships, as shown in Fig. 3.7.
It would be time consuming to describe the figure in detail, but interpretation is not difficult. For
example, fractured rocks fall above the sandstones because their porosity (fracture porosity) is very
low, yet these fractures form very connected networks that allow the efficient passage of fluids, and
hence the permeability is high. Such permeability may be directional because of preferred orientations of
the fractures. By comparison, clay cemented sandstones have high porosities, but the porosity is mainly
in the form of micro-porosity filled with chemically and physically (capillary) bound water which is
immobile. This porosity does not take place in fluid flow, so the permeability is low.
It might be expected that grain size also has some control on permeability. Figure 3.8 shows a
poroperm cross-plot for a well in a carbonate reservoir where the grain size, porosity and permeability
were measured for each core taken. Taking the data as a whole, there is little in the way of a clear
trend. However, trends emerge when the individual grain size fractions are considered. Now it is clear
that rocks with smaller grain sizes have smaller permeabilities than those with larger grain sizes. This
is because smaller grain sizes produce smaller pores, and rather more importantly, smaller pore
throats, which constrain the fluid flow more than larger grains which produce larger pore throats.
In summary, permeability:
• Depends upon porosity.
• Depends upon the connectivity of the flow paths in the rock.
• Depends, therefore, in a complex way upon the pore geometry of the rock.
• Is a directional quantity that can be affected by heterogeneous or directional properties of the pore
geometry.
Carbonate
100
10
1
0 10 20 30 40
Porosity, %
> 100 microns
50 - 100 microns
5 - 50 microns
<5 micron
Some of the common methodologies for estimating permeability from logs include:-
• Deriving a porosity/permeability relationship from core, and applying this to log derived
porosity over un-cored intervals.
• Acoustic/Sonic “Stoneley” wellbore waves respond to pore connectivity with the wellbore,
reducing the stoneley wave amplitude over permeable intervals.
• Formation tester tools draw a small sample of fluid into the tool, and using a modified darcy
equation give an estimate of “Drawdown” permeability. It is also possible with modern tools to
set a pair of inflatable packers over a zone and undertake a miniature well-test.
We have seen that the viability of a reservoir depends upon three critical parameters. The first two of
these are the porosity of the reservoir rock, which defines the total volume available for hydrocarbon
saturation, and the permeability, which defines how easy it is to extract any hydrocarbons that are
present. The final critical parameter is the hydrocarbon saturation, or how much of the porosity is
occupied by hydrocarbons. This, and the related gas and water saturations are controlled by capillary
pressure.
The pore space in a rock is occupied by fluids. In hydrocarbon reservoirs these fluids are hydrocarbon
gasses, oil and an aqueous brine. We define the pore fraction of each of these as Sg, So and Sw,
respectively. Hence, Sg + So + Sw = 1.
The amount of each of these fluids present at a given level in the reservoir depends upon gravity
(buoyancy) forces, which tend to stratify the reservoir fluids according to their density, external
hydrodynamic forces such as flow from a remote aquifer, and interfacial forces that act between the
various reservoir fluids and between the fluids and the rock matrix.
The interfacial forces either take the form of (i) forces on the interface between two fluids, or (ii)
between the fluid and the solid matrix of the rock. Both effects rely upon differences in the relative
strength of inter-molecular forces between gasses, liquids and solids.
• Liquid-gas or liquid-vapour forces result from the differences in molecular attraction of each
of the gas and liquid molecules for molecules of the same fluid compared to the molecular
attraction for molecules of the other fluid.
• The liquid-liquid forces result from the differences in molecular attraction of each of the liquid
molecules for molecules of the same liquid compared to the molecular attraction for molecules
of the other liquid.
• The fluid-solid forces result from the preference for fluid molecules to be attracted to the solid
mineral surface rather than to molecules of the same fluid.
The interfacial forces give rise to what is known as a capillary pressure. Capillary pressure is the
difference in fluid pressure across an interface between two fluids in a confined volume.
In the discussion of capillary pressure, we will first examine surface energy and wettability, and then
go on to derive a general expression for capillary pressure.
its fellow molecules inside the drop, but only weakly attracted to molecules outside the drop. This is
because there are much fewer molecules in the vapour phase that surrounds the drop. Hence, the
molecules on the surface of the drop experience net inward attraction. Figure 4.1 illustrates this
schematically.
This attraction ensures that the drop attains the shape with least surface area in any given environment.
For example, in zero gravity, the drop will float and will be perfectly spherical; the sphere being the
geometry with the least surface area per volume.
To summarize; in the drop there are attractive forces between the molecules. The molecules inside the
drop are surrounded by balanced attractive forces. The molecules at the surface experience a net
inward force, which results in the drop contracting to the smallest surface area. The fluid pressure
inside the drop is higher than that outside it, therefore the drop behaves as though it was contained in a
skin.
If we have a drop of liquid suspended in a gas, and we want to increase the surface area of the drop,
we must do work to move molecules from the interior of the drop to the surface. This is equivalent to
saying that the surface possess free surface energy.
F
F
F
δR
R
F F
2
πR
F F
Take the surface of any liquid droplet, and imagine that a circle is drawn on this surface with radius R.
The area of the circle is πR2 (Fig. 4.2). To increase the surface area of this drop by one unit of surface
area, work must be done to move molecules from the interior of the drop to its new surface. The
surface energy σ is equivalent to the work done W per unit surface area increase δA, i.e., σ =
W/δA. But the work done W is the force operating to make the change F multiplied by the distance the
force moves, so in the case of an imaginary circle on the surface of the liquid drop which increases
its diameter from R to R + δR, we can see that W = FδR and hence, σ = FδR /δA. When the
droplet expands the surface area of the droplet and the area of the circle also expand. If the change
in the radius of the circle is δR, the expanded circle has area π(R+δR)2, and the change in area has
been δA = 2πRδR (Fig. 4.2). Hence, σ = FδR/2πRδR = F/ 2πR = F/L p , where L p is the perimeter
of the original circle. Hence we can see that surface energy σ has the dimensions of
[force]/[length], and is therefore a form of tension. Hence, the surface energy of a fluid interface is
also called the surface tension of the fluid interface. Note therefore, that the surface tension is the force
operating perpendicular to a line on a fluid interface divided by the length of that line. In our case the
force was uniform around the circle, and the length of the circle (perimeter) was 2πR.
We can write:
σ = F (4.1)
L
The surface tension of a fluid-fluid interface is commonly measured by using a loop of wire. If loop
of wire is placed upon a fluid-fluid interface, it will be held in place by surface tension. Figure 4.3
shows the general arrangement. In this case we will be assuming that the lower fluid has the greater
forces of attraction between its molecules. If we apply a small upward force to the loop it tries to leave
the surface, but the surface tension hold the loop down (Fig. 4.3). There are two fluid interfaces that
contribute to this, the outer circle and the inner circle. Both need to be overcome to remove the loop
from the interface. The force required (from Eq. (4.1)) is F = F in n er + F ou t e r = 2πσR in n e r +
2πσR o u te r = 4πσR m e a n . Hence;
σ = F (4.2)
4πRmean
where, F = the force required to lift the loop from the surface interface (dynes or N)
σ = the surface tension (dynes/cm or N/m)
Rmean = the mean radius of the loop (cm or m).
2Rinner
2Rmean
2Router
Table 4.1 shows the surface tensions of some common fluid-fluid interfaces.
Note that the liquid drop cannot exist in isolation. The arguments put forward so far have considered a
single fluid in a gas. In practice both the fluid in the droplet and the fluid in which the droplet exists
contribute to the interfacial effects. It is the relative attractions of the molecules of each fluid for
themselves and each other that are important. In the case of a liquid drop in a gas, the effect of the gas
can usually be ignored. This is because the cohesion forces operating between the molecules in the
fluid are much greater than those in the gas because the fluid has a much higher density (number of
molecules per unit volume).
The interfacial tension of an interface is a fundamental property of that interface. If a liquid has an
interface with a gas, then some molecules of the liquid will be found in a vapour phase in the gas
above the interface. These molecules reduce the surface tension. As temperature increases the
concentration of the vapour phase increases, and the surface tension decreases. At the critical point
there are as many molecules in the gas phase as in the liquid phase, the interface effectively
disappears, and the surface tension becomes zero.
The wettability of a surface is the type of fluid which is preferentially attracted to that surface.
Figure 4.4a shows the general case of a single solid immersed in two immiscible fluids at an unstable
initial state. Now, imagine that the molecules of Fluid A are attracted to the solid molecules more than
those from Fluid B are. Fluid A will then displace Fluid B on the surface. However, the molecules of
Fluid A and Fluid B are contemporaneously also experiencing attractions from their fellow molecules
and, to a smaller extent, the molecules of the other fluid. Consequently, Fluid A cannot displace all of
Fluid B from the surface unless the attraction between the molecules on the surface of the solid and
those in Fluid A is very great. In the general case a stable state will be encountered where Fluid A
occupies the majority of the surface. At this equilibrium state, the contact angle θ characterizes the
balance between the various cohesive forces (Fig. 4.4b).
(a) (b)
Solid Solid
Figure 4.4 Surface wettability.
Clearly, the contact angle can be used as a measure of wettability. Figure 4.5 and
Table 4.2 show three possible scenarios.
Fluid B
(gas) Fluid A
(Water)
Solid (glass)
Fluid B
(gas)
Fluid A
Solid (glass)
Fluid B Fluid A
(gas) ____ (Mercury)
Solid (glass)
.
Figure 4.5 Contact angles for different wettabilities
.
Table 4.2 Contact angles for different wettabilities.
Contact Angle
relative to Fluid A, θ Description
(degrees)
0 Extremely Fluid A wet
0-30 Significantly A wet
30-60 Moderately Fluid A wet
60-90 Weakly Fluid A wet
90 Neutrally wet
90-120 Weakly Fluid B wet
120-150 Moderately Fluid B wet
150-180 Significantly B wet
180 Extremely Fluid B wet
If the sphere is increased in size by increasing its radius from R to R+δR, the increase in surface area
2 2 2
δA i s 4π (R +r δR +δR 2 ) -4π R , w h i c h as δR →0 i s δA = 8π R δR.
To increase the radius of the sphere from R to R+δR requires work to be done moving molecules from
the inside of the droplet to the surface. As work done W is the product of the surface tension σ, and the
change in surface area δA, we can say W = σδA = 8πσRδR.
But the work done is also the force operating to cause the change multiplied by the distance the force
moves. By definition any pressure P is equal to the force F operating per unit area A, perpendicular to
that area. Hence, F = PA. The force operating on the spherical surface of the droplet is therefore equal
to the nett pressure (Pin – Pout) multiplied by the area of the spherical droplet 4πR2, i.e., F =
4πR2(Pin – Pout). Since P in > P out , the net pressure (Pin – Pout). is positive, and the net pressure
and resulting force is outward. The distance the force moves is also outward , and its value is clearly
2
δR. Hence, the work done is W = FδR = 4πR (Pin – Pout)δR.
(4.3)
Note that the derivation equates the work done. This type of derivation does not rely on the interface
of the spherical droplet being static, as it is based upon the conservation of work (energy). It is also
independent of direction because work is a scalar quantity.
The capillary pressure is defined as the difference between the pressures in each of the two fluids
forming an interface. Hence Eq. (4.3) is an equation for capillary pressure.
(4.4)
Note that the pressure is proportional to the surface tension and cosθ , but inversely proportional to the
radius of the tube. Hence, as the radius of the tube increases, the capillary pressure decreases.
The difference in pressure (the capillary pressure) causes the interface to rise up the capillary tube
until the weight of the suspended column of fluid balances the capillary force that is associated with
the capillary pressure. Hence Eq. (4.4) can be used to describe the rise of fluids in a capillary tube.
A proportion of this force acts up the capillary tube F up = F cosθ = 2πR tube cosθ
At equilibrium (i.e., the meniscus is static), the upward capillary force is balanced by an equal and
opposite force due to the weight of the suspended column of fluids.
The weight of the suspended column of fluids is πR2 tube h g (ρliquid - ρvapour), where h is the height of
the fluid column, g is the acceleration due to gravity, and ρliquid and ρvapour are the densities of the
liquid phase below the meniscus and the vapour phase above the meniscus, respectively.
(4.5)
Hence, the height to which the fluid rises depends inversely upon the tube radius (pore throat radius);
the larger the tube (pore throats) the less the fluid rises.
The capillary pressure in a simple system of glass tubes of different radii can be calculated from Eq.
(4.5) by measuring the height of the capillary rise, while knowing g, and the densities of the liquid and
the vapour. In practice the density of the vapour is small at surface conditions and can be ignored. It
cannot, however, be ignored at down-hole conditions, where its density may approach the density of
the fluid.
Gas/Water System. A model of the capillary pressure in a gas/water system is shown in Fig. 4.7. The
free water level is the height of the interface when the radius of the capillary tube tends to infinity (i.e.,
the capillary pressure is zero and h = 0). In Fig. 4.7, this is approximated by a wide bath of fluid. The
interface at the free water level may exist at any given absolute pressure PFWL. Capillary forces exist
inside the restricted capillary tubing that result in the rise of the water to a height h above the free
water level.
The pressure in the gas phase above the meniscus in the capillary tube is P gas , where P gas =
P FWL − P gas g h.
The pressure in the water phase below the meniscus in the capillary tube is Pwater, where Pwater =
PFWL −ρwater g h.
Hence the capillary pressure is P cap = P gas – P wat er = (ρ water -ρ gas ) g h, which is consistent
with Eq. (4.5).
Using Eq. (4.4), the height to which the water rises can be obtained by rearranging this equation to
give
(4.6)
θ
Pgas
Pwater Gas
h
PFWL
Water
Oil/Water System. A model of the capillary pressure in a oil/water system is shown in Fig. 4.8. The
free water level is the height of the interface when the radius of the capillary tube tends to infinity (i.e.,
the capillary pressure is zero and h = 0) as before. This interface exists at a given absolute pressure
PFWL. Capillary forces exist inside the restricted capillary tubing that result in the rise of the water to a
height h above the free water level.
The pressure in the oil phase above the meniscus in the capillary tube is P oil , where P oil = P FWL
− ρ oil g h.
The pressure in the water phase below the meniscus in the capillary tube is Pwater, where Pwater =
PFWL − ρwater g h.
Hence the capillary pressure is P cap = P oil –P water = (ρ water -ρ oil ) g h, similarly to the last case.
Using Eq. (4.4), the height to which the water rises can be obtained by rearranging this equation to
give
(4.7)
Pgas
θ
Pwater Gas
h
PFWL
Water
Knowledge of Parameters. The capillary pressure depends most critically upon the interfacial
tension and the wetting angle. These parameters change with pressure and temperature for any given
fluid/fluid/solid system. Measurements can be made accurately in the laboratory in simplified systems
such as the capillary tube, but are very difficult to carry out in the rock itself. The result is that we use
a range of values which are generally considered to be standard, but should not be considered to be
accurate. These are given in Table 4.3.
Table 4.3 Standard interfacial tension and wetting angle values for common mixtures of reservoir
fluids and solids.
Capillary h
Pressures
FWL
Fluid B
Fluid B Pressure
Line
Pressure
Figure 4.9 Capillary rise and capillary pressures.
The capillary pressures are shown in the graph in Fig. 4.9. We can see that Fluid A (gas) has
increasing pressure with depth due to the column of Fluid A above it, and this is described by the line
labeled ‘Fluid A Pressure Line’, which has a gradient ρFluidA g h. Fluid B (water) has increasing
pressure with depth due to the column of Fluid B above it, and this is described by the line labeled
‘Fluid B Pressure Line’, which has a gradient ρFluidB g h. Fluid A sits above Fluid B because it is less
dense. Its lower density results in it having a steeper gradient. Where the two line intersect, the
pressures in both fluids are the same and the capillary pressure is therefore zero. This is the free water level.
Figure 4.11 A force is needed to overcome the capillary pressure to displace a wetting fluid with a
non-wetting fluid.
If we have a capillary tube or a rock that contains 100% saturation of a wetting fluid (e.g. water) and
we introduce a non-wetting fluid (e.g. oil) to one end, the non-wetting fluid will not be spontaneously
drawn up into the tube or the pores of the rock. This is because the wetting fluid is held inside the tube
or the rock by the capillary force. To introduce the non-wetting fluid into the tube or the rock, we must
apply an external force to overcome the capillary force holding the wetting-fluid in place (Fig. 4.11).
This is called the displacement force or displacement pressure, and is equal in magnitude, but opposite
in sign (direction) to the capillary force and capillary pressure respectively.
IMPORTANT: In oil reservoirs, the reservoir rock initially contains water and is water-wet. Oil
migrates into the reservoir rock displacing the water. A displacement force is required to overcome the
capillary forces in the water saturated water-wet reservoir rock. This force is supplied by gravity
operating upon the differential buoyancy of the two fluids that result from their different densities.
There is a level at which oil cannot replace water further because the driving force is insufficient to
overcome the capillary force. This does not occur at the free water level, but at some height above the
free water level defined by the displacement pressure. Thus, in reservoirs, the oil/water contact (OWC)
is above the free water level (FWL).
Hence, if we have a number of capillary tubes of different radii arranged horizontally and containing a
wetting fluid (e.g., water) (Fig. 4.12), and we wish to force a non-wetting fluid into the tubes (e.g.,
oil), a lower displacement pressure will be required to replace the water in the larger tubes than to
replace the water in the thinner tubes. One can use the analogy with rocks, but remember that the pore
space in rocks is composed of pores of different sizes connected by pore throats of varying size that
are smaller than the pores. Hence the displacement force or pressure required to force a non-wetting
fluid into a rock saturated with a wetting fluid will be controlled by the size of the pore openings.
Oil
Oil
Oil
Water
Water
Oil In
An Example. Assume that a rock is composed of five sizes of pore in equal number, and for the sake
of simplicity, these pores are connected by pore throats of the same size as the pores. The rock is
completely saturated with a wetting fluid. Table 4.4 shows the size, number and volume of pores
together with their respective displacement pressures.
The data given in Table 4.4 can be plotted as a displacement versus pressure plot (Fig. 4.13). In this
rock the pressure required to begin to displace water by oil is the initial displacement pressure (2800
Pa in this case). The level in the reservoir at which this is the case is called the 100% water level
which should be the same as the OWC, and is above the FWL by a height related to the size of the
displacement pressure and controlled by the largest pore openings in the rock. If the rock was
composed of 100% of pores of 25 micron size, this would be the level above which there would be
100% oil saturation and below which there would be 100% water saturation. However, the rock also
contains smaller pores, with higher displacement pressures. Thus there will be a partial water
saturation above the 100% water level occupying the smaller pores, and this water saturation will
reduce and become confined to smaller and smaller sized pores as one progresses to higher levels
above the 100% water level. Water will only be displaced from a given pore size if there is a
sufficiently large force to overcome the capillary force for that size of pore. Commonly, the force
driving the oil into the reservoir rock (due to differential buoyancy) is insufficient to overcome the
capillary forces associated with the smallest pores. Hence, the smallest pores in an oil zone remain
saturated with water, and there is an irreducible water saturation (Swi) in the reservoir.
80
60
0.13%
40
0.82%
20
3.3%
13.2% Initial
82.5% Displacement
Pressure
0
0 25 50 75 100
Water Saturation (%)
Figure 4.13 Displacement/pressure curves.
Note that once oil is in place, the operation of temperature and pressure may geochemically alter the
properties of the mineral surfaces such that they become oil-wet. This occurs only in the presence of
oil, so it can be imagined that a mature reservoir may have a mixed wettability, with water-wet small
pores containing water and oil-wet larger pores containing oil. This has implications when we want to
move oil and water through the rock during production.
2000 80
2100 60
Oil
Gradient
2200 40
Water
Gradient
2300 20
FWL
Initial
2400 Displacement
0
0 25 50 75 100
Pressure
0 25 50 75 100
Fluid Pressure (psi) Water Saturation (%)
Water
Saturation
Water Saturation
Capillary Pressure versus Height
(Displacement 2000
Pressure)
gives height
above FWL
2100
for a given water
saturation,
h = Pc / ∆ρg
2200
100% Water
Height of the FWL
2300 Level
FWL
2400
0 25 50 75 100
The curve shown in Fig. 4.13 is continuous for a reservoir rock because the rock contains a continuous
pore size distribution. This curve is called the capillary pressure curve. The capillary pressure curve
describes the pressure required to displace from the rock a wetting fluid at initially 100% saturation to a
given saturation.
The capillary pressure is related to the height above the FWL by the relationship
Pc = ∆ ρg h (4.8)
So, if the capillary pressure curve is known, and the FWL is known, we can calculate the water
saturation at any point in the reservoir as in Fig. 4.14. If this water saturation agrees with the water
saturations obtained from wireline tools and core in a given well, the wireline water saturation data can
be used with confidence as a predictor of water saturation in the uncored sections of other wells in the
same field.
Capillary
Gas Pressure
Cor e Plug
1 of 10 per Brine Outlet
Apparatus
32
16
0
0 25 50 75 100
Water Saturation (%)
The maximum pressure is limited by the pressure at which the porous plate begins to let the gas or oil
pass through with the water.
There is also a range of specialized dynamic and semi-dynamic versions of the porous plate technique
that can be used on confined cores together with direct measurement of sample fluid saturations using
X-Ray, CT and gamma radiometric imaging techniques.
4.6.2 Centrifuge
The sample is saturated completely with a wetting fluid (usually water), and its effective porosity is
measured. It is placed in a centrifuge and rotated at progressively higher speeds. The speed of rotation
generates a centripetal force that displaces the wetting fluid from the sample replacing it with air or oil
(if an oil reservoir is attached to the inside of the rotating sample). The volume of the evolved fluid is
measured by collecting it in a graduated vial attached to the outside of the spinning arrangement, and
read using a stroboscope. At slow rotation speeds, the force is only sufficient to displace water from
the largest pores. At higher speeds the force is able to displace water from smaller and smaller pores in
the sample. The pressures on the fluids can be calculated from an equation based upon the rotary
motion of the centrifuge. This is plotted against the water saturation in the sample calculated from the
measured volume of evolved fluids for each spin speed (Fig. 4.17).
40
30
20
10
0
0 25 50 75 100
Water Saturation (%)
Advantages: Fast.
Disadvantages: Commonly only used with air as the displacing fluid.
Only provides a few data points on the capillary pressure curve.
As pressure is a function of distance from the spin centre, the fluid saturation
across the sample varies, and the measured values are taken as the mean value.
Pressure
Vessel
Pressure
Oil In
Drop in
Hg Level
Pressure
Oil
Drop in
Glass Hg Level
Holder
Sample
Dried Sample is Mercury is added until mark Atmospheric pressure is Pressure is raised to 200 MPa
Evacuated while the system remains a p p l i e d t o t h e s ys t e m allowing Hg to enter progressively
evacuated. The amount of allowing mercury to enter smaller pores. The drop in the Hg
mercury needed provides a the larger pores. Drop in Hg level gives the volume of the pores
measurement of bulk volume level gives volume of saturated with Hg at each pressure
macropores. level.
Advantages: Fast.
Provides porosity of sample.
Provides pore size distribution.
Provides grain size distribution.
Can be used on small, irregular samples, such as cuttings.
Gives many data points for intrusion.
Disadvantages: Cannot be used with reservoir fluids.
Data must be corrected to give reservoir values.
Can only be used in imbibition cycle.
We require the capillary pressure curve for the reservoir fluids and the reservoir rock. Measurements,
however, are generally made with other fluids. We must, therefore have a method for correcting the
data to give the capillary pressure for the reservoir fluids. The corrections are simple and take account
of the different interfacial tensions and wetting angles for different fluid combinations at 20oC only.
(4.9)
(4.10)
In both cases the values for interfacial tension and wetting angle come from Table 4.3. Since the value
of R is the same whatever the fluids are, we can equate and rearrange Eqs. (4.9) and (4.10) to give
These conversions are valid for the values of interfacial tension and wetting angle measured at 20oC
and given in Table 4.3. However, these parameters change with pressure and temperature, and are
therefore different in the sub-surface. Values of interfacial tension and wetting angle are very difficult to
obtain for sub-surface conditions, and one might commonly have to make do with those at 20oC.
However, sub-surface values should always be used when available as the interfacial angle is
particularly sensitive to changes in temperature, with values as low as 10 dynes/cm for oil/water
systems, and as low as 25 dynes/cm for gas water systems.
Many of the values in Table 4.5 are given to 4 significant figures accuracy. However, it is common for
only two significant figures to be used (i.e., 10 for 10.5 and 5 for 5.1). This is often justified on the
grounds that our knowledge of the values of interfacial tension and wetting angle is not accurate. The
lack of accurate data introduces errors in any final calculations. However, noting that these
inaccuracies occur is no excuse for introducing further voluntary inaccuracies by making such
assumptions as using 10 instead of 10.5. The error introduced by this assumption can result in errors of
millions of dollars on the bottom line.
Now we can convert mercury derived capillary pressure data to the capillary pressure for the reservoir
fluids, we can write Eq. (4.8) in terms of the mercury capillary pressure.
(4.14)
and
(4.15)
These equations are very useful, because they allow the height in the reservoir for a given saturation to
be calculated from capillary curves measured by mercury porisimetry, which is the most common
capillary pressure technique practiced in the industry.
In deltaic-type deposition sands may be laid down in fan-like bodies. The current that is responsible
for the deposition is fast in the proximal region and becomes progressively slower as one progresses
distally or laterally. Hence, in the proximal region the grain size of the resulting rocks is large (only
large grains have time to fall from the fast current). Distally and laterally the grain size of the resulting
rocks becomes finer as smaller sand particles have time to fall out of the slower current. Eventually, at
the extreme edges of the fan, the current is so slow that very fine clay particles are also deposited,
becoming interspersed with the fine sand grains.
• The proximal large grained clean sands have large pores and large pore throats. Therefore, they
have small initial displacement pressures, and their water can easily be displaced by migrating oil.
• The distal fine grained clean sands have small pores and consequently small pore throats. They
have higher initial displacement pressures, and their water cannot be displaced by migrating oil as
easily as for the coarser grained rocks.
• The distal fine grained shaly sands (containing interspersed clay) have the smallest pores and
consequently the smallest pore throats, and these pores and pore throats are blocked by the clay
particles. They have very high initial displacement pressures; the great majority of their water is
immobile and cannot be displaced by migrating oil. Figure 4.19 shows the results of 5 wells drilled
progressively more distally in the same sub-marine fan.
Well D Well B
Well C
Well E Sediment
Supply
Proximal
Capillary Pressure
Curves Shaliness
50% Water
Level
100% Water
Level
0 20 40 60 80 100
Water Saturation (%)
Shaliness
The first part shows the fan geometry and position of the wells (labeled A to E). The second part
shows the capillary pressure curves for each of 5 samples taken from each of the wells with the 100%
water point labeled S100A to S100E, the initial displacement pressure labeled DIA to DIE, the 50%
water point labeled as S50A to S 50 E, and the irreducible water saturation labeled S wi A to S wi E.
The third part shows the height above the free water level as a function of facies and water saturation, with
the same saturation labels as in part 2. Note the following:
• The initial displacement pressure becomes greater as the sample becomes more shaly (finer
grained) as one moves distally.
• The irreducible water saturation becomes greater as the sample becomes more shaly (finer
grained) as one moves distally.
• The capillary pressure curve for the proximal samples are quite flat from 100% water to about
60% water indicating that only low pressures are required to replace about the first 40% of water
with oil (in the largest pores with the largest and most accessible pore throats).
• At any given capillary pressure, more water is retained as the sample becomes more shaly (finer
grained) as one moves distally.
• The height above the free water level of 100% water saturation increases as the sample becomes
more shaly (finer grained) as one moves distally.
• The height above the free water level of 50% water saturation increases as the sample becomes
more shaly (finer grained) as one moves distally.
So, the grain size controls the height of the 100% water saturation level. Now imagine four layers of
sand that have different grain sizes, porosities and permeabilities, and that have become folded (Fig.
4.20). Each layer has its own capillary pressure characteristics associated with its grain size, and that
capillary pressure is related to the height of the 100% water level above the free water level as shown. It
can be seen that it is possible to drill through layers which are 100% water zones existing between oil
or gas zones from the same reservoir. In this case Well C has two oil water contacts. It is the bottom
one that is the true OWC. Because multiple OWCs are confusing, we tend to talk about oil down to
(ODT) and gas down to (GDT) levels when there are multiple hydrocarbon/water contacts.
5. LOGGING
An example of a more usual log is given as Fig. 5.2. This log contains 3 tracks and many different sets
of log data. There is a very large range of basic logs that can be run in a well, and an even greater
number of derived logs, i.e., data that can be plotted in log form that is calculated from the basic logs.
Examples of the basic physical parameters that can be measured down-hole with logs include (a) the
size of the borehole, (b) the orientation of the borehole, (c) temperature, (d) pressure, (e) the natural
radioactivity of the rocks, (f) the acoustic properties of the rocks, (g) the attenuation offered by the
rocks to radioactivity generated from the tool, (h) the electrical properties of the rocks, (i) the NMR
characteristics of the rocks, and so on.
Examples of derived logs include (a) porosity derived from the acoustic/sonic or density log, (b)
water saturation calculated from the porosity and the electrical logs.
Wireline logs are made when the drill-bit is removed from the borehole. This can be either between
drilling episodes, before casing is laid, or at the end of drilling. Wireline logging can be done when the
newly drilled rock formations form the wall of the borehole (open-hole logs) or after a concrete lining
or casing has been inserted to stabilize the well bore (cased-hole logs). However, as may be guessed,
the quality of data from the rock is best from open-hole logs, and some measurements cannot be done
at all through casing. Cased-hole logging requires special tools to optimize measurement through the
casing, and is a specialist subject not covered by this course.
A wireline logging set-up consists of the wireline tool, the wire itself, a winching gear, and a vehicle
containing data analysis and recording equipment. The vehicle is usually a large lorry (Fig. 5.3) for
land-based measurements, or a transportable cabin for use on rigs (Fig. 5.4). All data recording and
processing is now done digitally using powerful computers and array processors (Fig. 5.5).
The required tool is attached to the wire and lowered into a cleaned and stable borehole from which
the drill string and bit has been extracted. When it reaches the bottom of the interval to be logged, it is
slowly withdrawn at a predetermined speed. Log measurements are made continually during this
process. The wire is made from a bundle of cables, the outermost ones of which are steel and carry the
load, and the innermost of which are insulated copper and carry the electrical signals. The wireline is
winched up the borehole by a motorized drum typically at around 30ft/min (9m/min) (1800 ft/hr,
540m/hr). It should be noted that most logs are run during their withdrawal from the borehole, which
gives greater cable tension/depth control. One exception is the temperature log, which must be run on
the way into the borehole because its presence in the hole disturbs the temperature reading.
The tension in the wire and the speed of the wire is constantly monitored so that intermittently stuck
tools can be detected. This information is used together with knowledge about the natural elasticity of the
wire that causes it to stretch under its own weight to calculate the instantaneous depth throughout the
logging phase.
Data is transmitted to the surface through the wire and recorded digitally. Modern tools sample the
formation at 6” (0.1524m) intervals, or far higher for the high definition imaging tools, so for a
combination tool with several different logs, there may be several 100,000’s of individual
measurements over a log run. Depending on the logging speed, data transmission rates may be as high
as 200 kbits per second. This data is subjected immediately to a rudimentary analysis by powerful
processors and displayed so that the quality of the measurements can be checked. The data is recorded onto
hard disk and digital (DAT) tape and can be transmitted by satellite onto the (secured) Internet for in-
depth analysis anywhere it is required.
A typical drilling and logging history for a well might be as shown in Table 5.1. Note that the logs are
run in each section before casing takes place. When each log is run, the operator ensures that some
measurements are included opposite the previously cased section above the log run as a test that the
log tools are working correctly and to provide a local depth fixed point to help tie the depths for each
tool together. Whenever a borehole is logged in sections the operator also ensures that there are
overlap sections so that the logs can be tied together accurately, and repeat sections to check log
quality and consistency.
Depth
Procedure Drill Bit Size Casing Diameter Tools
(m sub sea/
(inches) (inches)
surface)
Drill 25 - 300-400
Case 20 300-400
Drill 17.5 - 400-700
Log - - 700-400 Res-AC-GR/1
Case - 13.375 300-690
Drill 12.25 - 700-1400
Log - - 1400-680 Res-AC-GR/2
Log - - 1400-680 Den-CN-GR/1
Log - - 1400-680 Dip-GR/1
Case - 9.625 690-1390
Drill 8.5 - 1400-2100
Log - - 2100-1380 Res-AC-GR/3
Log - - 2100-1380 Den-CN-SGR/2
Log - - 2100-1380 Image/1
Log - - 2100-1380 FT/1
Note that the most tools are run in the area that is expected to be of greatest interest (i.e., the reservoir).
For tool codes see wireline operators manuals.
Efforts are made at all times to ensure that the logs are of the highest quality. Methods of ensuring that
logging quality is high include (a) making repeat sections of log measurements over short sections that
contain both low and high measured values, (b) making overlap measurements of the previous
(shallower) run of the same tool, (c) making measurements opposite the casing at the top of the run,
(d) local calibration of all logging tools both before and after the survey, and (e) ensuring that a
‘master’ calibration of the tools has been carried out recently at the wireline service company’s
workshop (particularly for sensitive and important ‘porosity’ tools).
LWD data is recorded by instruments which are part of the drillstring. This means that formation
Geoff Page 2010 Page 60
Petrophysics MSc Course Notes Logging
measurements are acquired real-time while the hole is being drilled, providing continuous updates of
the formations that are encounteencountered. These measurements can be used to steer the wellbore
through target reservoir zones, maximizing the production potential – a process known as “Reservoir
Navigation”. For LWD measurements there is no “cable” connecting the instruments to the surface for
the transmission of signals and measurement of depth. Depth is measured from the length of drillpipe in
the wellbore, the instruments are powered by a turbine generator from the mud flow, and a basic set of
data is sent to the surface as a series of pressure pulses in the mud inside the drillstring. This data
transmission system is typically only 5-10 bits/second (very low compared to Wireline), however it is
sufficient to transmit the required basic log curves and a full dataset is also stored in downhole memory
in the tools for retrieval at surface when the drillpipe is pulled out.
The decision on which type of measurements are acquired is a combination of economics (LWD tools
are rented for the entire drilling time, but Wireline tools need extra rig time after drilling is completed),
the availability of sensors (Wireline are generally more comprehensive measurements), and the type
and urgency of the data required.
Track 1, historically known as the “P” (Potential = SP) track, on the leftmost side is always linear and is
often reserved for drill bit size, caliper and gamma ray tool information.
Moving to the right, there is a depth column that may be in feet or metres. A range of depth scales are
used depending upon the resolution required for a particular analysis. In Europe and the East scales of
1:1000 and 1:500 are often used for large scale correlations between wells, while 1:200 is optimum for
petrophysical evaluation, 1:40 for correlation with core information, and 1:20 for detailed dipmeter or
image analysis. In the USA they use 1:1200. 1:600 and 1:240. However, the digital recording of logs
now ensures that they can be displayed and printed at any required scale. Scales may also be defined in
inches of paper/100ft of wellbore i.e. a 6” plot is equivalent to 1/200, 1.2” is equivalent to 1/1000.
Moving once more to the right, there are tracks 2 and 3, historically known as the “R” (Resistivity)
track, which may have either a linear or a logarithmic scale and may be combined into one. These
tracks traditionally were reserved for density, porosity and electrical tool data.
All tracks can take multiple log curves, and the code for the curve, its style (i.e., green dashed line),
and the scale units are given in a stack at the top of the log. The curves are arranged in such a way that
they can wraparound (i.e., if the curve slips of the top of the scale it appears again at the bottom, and
the reader must make the required scale adjustments to read the wrapped-around or “backup” data).
The fundamental basis of the API standard is still used, but the advent of computer analysis has made
the presentation of logs rather easier. The modern log data is not only presented as hardcopy, but will
come to you as an e-mail, on a tape or CD-ROM. The arrangement of this data in the form of logs on a
display or a printout is now likely to have as many tracks as needed (its up to you!). Each will be linear
or logarithmic, or even use a core image or a coloured array to represent the data. However the depth
track is usually retained between tracks 1 and 2, and may be repeated further over for very wide plots.
Electrical data is almost always presented on a logarithmic track because it can vary over many orders
of magnitude.
There are numerous little markers and ‘grace notes’ present on logs that require to be learnt. The
The main section of logs will have a comprehensive header sheet that contains a plethora of
information about how the log was run. Some of this information also appears at the end of the logs,
together with repeat or overlap sections and information on tool calibration.
Figure 5.6 Log presentation grids. Note that the API format defines particular dimensions for each
track, and an overall width of 8.25 inches.
5.4 Tools
There are a plethora of logging tools now available in the oil industry. There can be split into the
following types:
• Production tools
These are specialist tools for running during the production of hydrocarbons for measuring
formation and wellbore fluids, pressure, flow rate, and are always of the slim-hole variety to
facilitate their insertion into a producing well under pressure.
• LWD tools
These tools are designed to connect into the drill string and communicate data to the surface
through the drilling mud. Their diameter is related to the drillstring into which they are built.
A summary of some of the main standard open-hole tools is given in Table 5.2.
Drilling rigs are expensive and can often amount to a high percentage of the total drilling costs. The
time required to run wireline logs effectively delays the drilling operation leading to greater expense.
To minimize this expense tools are bolted together for lowering down the borehole. All wireline tools are
designed to be combined in this way and there are many possible combinations, some of which are
as long as 60 m. Each tool in the combination has its own measurement-sensitive region, which
measures a given formation at a specific depth at different offset times. The data has to be depth
shifted to account for this. This is done in the data acquisition system with information about the
instantaneous speed and dimensions of the tool combination. Even using combinations, it is common for
as many as three log runs to be made in the same section of the borehole with different combinations
of tools. If this is the case, one tool is included in all combinations to act as a reference to which the
others might be compared for depth matching purposes. This tool is commonly the gamma ray tool
(GR).
LWD tools can also be combined as there is only one opportunity to record data real-time while
drilling. It is also possible to record data using LWD tools as part of a later trip into (or out of) the hole
with the drillpipe. These are sometimes called measurement after drilling or “MAD” logs!
Figures 5.7 and 5.8 show standard open-hole density logging tools, both Wireline and LWD. The
Wireline tool has a motorized arm which opens up downhole to push the measurement surface (pad)
against the rock. This arm also provides a borehole diameter measurement. The LWD tool has
ultrasonic Acoustic Stand Off (ASO) pulse-echo hole size measurements.
On land, where the general geology is well known and drilling is cheaper, only a few logs may be run.
This is because expense can be spared as information concerning the subsurface is not hard to come
by. However, and perhaps counter-intuitively, offshore a full set of logs is often carried out even when
there are no hydrocarbon shows. This is because drilling is so expensive that relatively fewer wells
will be drilled, and hence the maximum data must be squeezed out of each well regardless almost of
the cost. If there are no hydrocarbon shows, the well data will still be very useful for correlation with
other wells that do show hydrocarbons, will help to constrain the extent of neighbouring reservoirs, as
well as adding to the information about the general geological structure of the area.
Magnetometer
Tri-axial
ASO
Acoustic
Package
Long Spaced
Detector
Measurement
Surface
Short Spaced
Detector
Source Port
Stabilization
Logging conditions
Temperature Borehole temperature for resistivity
Temperature (BHT) Corrected with Horner plot
calculations.
Pressure (PRESS) Fluid pressure Fluid pressure for formation volume Incorporated in FT
factor calculations.
Caliper (CAL) Borehole diameter Data quality, in situ stress tensor, Available in 1, 2, 3, 4, or
lithology and permeability indicator multiarm versions.
Lithology
Gamma Ray (GR) Natural radioactivity of Shale indicator and depth matching Can read through casing.
the formation.
Density (FDC, Bulk density of the Total porosity Used to calculate synthetic
LDT, ZDL) formation. seismograms.
Total porosity +/- (shale
Neutron (SNP, Hydrogen Can read through casing.
concentration in the increases measured porosity, gas
CNL)
formation. reduces measured porosity)
Resistivity
Resistivity of flushed, Used in water saturation calculations. Now obsolete, not focussed,
Simple electric log
can’t be used in oil based
(SN, LN, Lat) shallow and deep zones
muds, prone to invasion.
respectively.
Focussed devices.
Induction Logs Conductivity of the Conductivity and resistivity in oil
Use in oil based and fresh
(IES, ISF, DIL, formation. based muds, and hence calculation of water muds. Range of depths
water saturation. of investigation.
PIL, AIT, HDIL)
(Vertical resolution 1-10 ft.)
6.1 Introduction
Logging has a single clearly defined purpose: to give accurate and representative data on the physical
properties of the rock formations and fluids encountered in a borehole.
The tools used to take these readings have to cope with extremely tough conditions downhole,
particularly, high temperatures and pressures, inhospitable chemical conditions and the physical
constraints imposed by the physics of the measurements and the borehole geometry. It should also be
remembered that we are interested in the properties of the rocks in undisturbed conditions, and the act
of drilling the borehole is the single most disturbing thing that we can do to a formation.
The pressure that a rock is subjected to at a given depth is determined by the weight of the rock above
it, and hence the density of that rock. This is called the overburden pressure or sometimes the
lithostatic pressure (note that, to a first approximation, the overburden pressure is the same in all
directions (isotropic)). We can write an equation to describe the overburden pressure
Clearly the rock above a given depth will have a varied lithology and porosity and hence a varying
density. A more accurate determination of the overburden pressure can be obtained by summing the
pressure contributions for each density by writing Eq. (6.1) for i different rock densities, each with
thickness hi.
(6.2)
The overburden pressure/depth curve is called the geobar or lithostat. An example is given in Fig. 6.1.
It should be noted that in actuality this pressure is not isotropic but operates vertically. The pressures
horizontally depend upon the overburden pressure, but are modified by additional large scale sub-
horizontal tectonic forces (in tension and compression), and are affected by local inhomogeneities in
the crust, such as fractures. However, to a first approximation the pressure at depth can often be
considered to be isotropic (hydrostatic).
Clearly the fluid in the rock above a given depth may have a varying density, hence a summed
equation may be better at describing the fluid pressure at depth. The connected fluid pressure is called
the hydrostatic pressure and is shown in Fig. 6.1.
Clearly, the fluids will not always be in good connection with the surface, and indeed this will
generally not be the case in reservoirs, where a seal is required to stop the hydrocarbons escaping. In
these conditions the fluid pressure may be much higher than that predicted by Eq. (6.3). These fluids
are called overpressured fluids (Fig. 6.1). Overpressurization is caused by a trap to stop fluids
escaping, and a process that will raise the fluid pressure such as compaction of the rock.
However, the fluid pressure is also affected by other forces. These are capillary pressure, buoyancy
forces (if the fluid is dynamically driven by a density gradient), and externally imposed fluid flow
forces from aquifers.
The extraction of hydrocarbons in a reservoir ultimately lowers the fluid pressure, and hence raises the
effective pressure experienced by the rock. Thus there may be occasions where, what was once stable
reservoir rock, now is exposed to a greater effective pressure, and begins to crush and compact. This is
currently occurring in several reservoirs in the North Sea, and causes huge problems for wellbore and
platform stability. Specialist logs can measure this compaction very accurately.
If a well could be drilled without a drilling fluid, formation fluids, which are under their fluid pressure,
would spurt out of the borehole (blow-out). The density of the drilling fluid used in a particular
borehole is designed to generate a drilling mud pressure (due to the weight of the drilling mud in the
borehole above a given depth) that counteracts the fluid pressure in the formation and prevent blowouts.
This is usually successful, but because sudden increases in overpressure can be encountered the drilling
mud pressure is kept higher than necessary as a safety feature. When the mud pressure is greater than
the formation fluid pressure, the well is said to be over-balanced.
Note that varying the drilling mud weight also helps to protect the wellbore from crushing by the high
lithostatic forces, especially in deviated and horizontal wells.
The drilling mud is a suspension of mud particles (a slurry) in an aqueous or oil-based medium. Each
of these will be discussed in more detail later in this section.
6.6 Invasion
6.6.1 Introduction
The drilling mud is at a pressure greater than the fluid pressure in the formation. When the drilling
mud encounters a porous and permeable formation, the drilling mud will flow into the formation under
the influence of this difference in fluid pressures. This is called invasion. However, the particulates in
the mud will be left at the surface, with the rock acting as an efficient filter. Thus, there is a build-up
of mud particles on the inner wall of the borehole, and this is called the mud cake. The remaining
liquid part of the drilling mud enters the formation, pushing back the reservoir fluids. This part of the
drilling mud is called the mud filtrate. The zone where the mud filtrate has replaced the reservoir
fluids is called the flushed zone and there is a zone further into the rock where the replacement of
reservoir fluids with mud filtrate is incomplete, which is called the transition zone. The Virgin
reservoir fluids occupy the uninvaded zone further into the formation.
Note that invasion only occurs for porous and permeable formations, and is a self limiting process
because the mud cake is an efficient block to further filtration and flow. However, there is a qualitative
relationship between low porosity and high permeability, and larger depth of invasion and thicker mud
cake. High permeability formations admit the mud filtrate easily, so the invasion is deep and the
mudcake builds up quickly to thick layers. High porosity formations allow the storage of more mud
filtrate per invasion distance, therefore the depth of invasion is smaller than for low porosity
formations of the same permeability.
The different fluid types in each of the zones, and hence the formations that they occupy, will have
different physical properties. Only the deepest zone contains the circumstances of the undisturbed
rock, while the Wireline or LWD tools occupy the borehole. Thus, to get good readings of the true sub-
surface properties of the rock, the tool has to either, (i) measure accurately through the borehole mud,
mud cake flushed zone and transition zone, or (ii) make readings closer to the tool (i.e., in the flushed
zone) that can be reliably corrected to represent the values in the uninvaded zone.
All logging companies provide correction graphs for their various tools. However, the accuracy of the
correction diminishes as the diameter of the borehole, the thickness of mud cake, and the depth of
invasion increase. Effort is therefore made to ensure that these are minimized by (i) ensuring the mud
chemistry does not create wash-outs increasing the borehole diameter, (ii) ensuring that the mud is not
so low salinity that it causes the swelling of formation clays, (iii) minimize mud filtrate, and (iv) set
the mud weight such that the mud pressure is only slightly greater than the formation pressures (only
slightly over-balanced).
One simple method of checking for invasion is by measuring the loss of drilling fluids. This is done
regularly at the wellsite for water-based drilling muds by forcing samples of drilling mud through filter
paper with a 100 psi pressure for 30 minutes and measuring the volume of the mud filtrate. For oil-
based muds a high temperature apparatus is required as the filtration of oil-based muds is low at
surface temperatures. While mudcake is a useful indicator of porous and permeable formations, it can
cause the drill-bit, drill-pipe and wireline tools to stick in the hole, and the associated deep invasion
causes problems for accurate log data acquisition.
dtrans
Invasion dflush
Diameters dhole
Shallow Invasion
Medium Invasion
Rxo Rt
Rmf Rm Rmc Rw Deep Invasion
Sxo
Sw
Rt
Rmc
Rm
Rxo Sxo
Rmf Sw
Rw
Figure 6.3 Effect of time and permeability on invasion depth for water-based muds.
In water-bearing formations the mud filtrate replaces all of the formation water close to the borehole
(Fig. 6.4) and this decreases with depth of invasion. In oil-bearing formations the mud filtrate replaces
all the formation water and most of the oil close to the borehole wall, again decreasing with distance
into the formation (Fig 6.4).
Formation
Water
Flushed Transition Uninvaded Flushed Transition Uninvaded
Zone Zone Zone Zone Zone Zone
Invaded Invaded
Zone Zone
Figure 6.4 Invasion profiles for water-based muds in water and oil bearing formations.
Oil-Based Formation
Oil-Based
Mud Formation Oil
Mud
Filtrate Water
Filtrate
Formation
Water
Figure 6.5 Invasion profiles for oil-based muds in water and oil bearing formations.
Drilling oil-bearing formations with salt saturated water-based muds. The high density salty mud
filtrate accumulates at the bottom of the formation resulting in a resistivity profile that has lower
resistivities at the base of the formation than at the top.
Drilling water-bearing formations with oil-based drilling muds. The low density oil-based mud
filtrate accumulates at the top of the formation, again resulting in a resistivity profile that has lower
resistivities at the base of the formation than at the top.
These effects usually disappear with time after the casing is set and cemented, remaining only if the
permeability of the formation is very low.
Depth of Investigation. Most tools have a shallow depth of investigation, resulting in the
measurement of the formation in the flushed or transition zones. Most radiation tools have an
investigation depth of less than 20” (0.5m) because they rely on measuring natural or induced
radiations that are adsorbed by the rock or the fluids they contain. Electrical tools come in various
versions with a wide range of investigation depths, from imaging tools which measure only a fraction of
an inch (a few centimetres), to the deep penetration tools 10-20ft (up to 6 m). The depth of
investigation often depends upon the density/porosity of the formation.
Vertical (Bed) Resolution. This is closely related to the depth of investigation and to the logging
speed. In general a tool with a large depth of investigation, has a low vertical resolution. The true
parameters of the rock formation can only be measured if the bed is larger than at least the source-
detector distance of the tool. Many modern logging devices use a combination of shallow (high
resolution) and deep (low resolution) measurements, which are combined mathematically to extract
the best characteristics of each in a process know as “synthetic focusing”. This results in
measurements with a high resolution but deep depth of investigation. If the beds are thinner than the
bed resolution for the tool, then the measured parameter is underestimated. This is the thin-bed effect.
Investigation Geometry. This depends upon the tool. Some tools have a pseudo-hemispherical zone
of sensitivity (the radiation logs), some toroidal (the induction logs), some tubular (acoustic logs), and
some focussed planar (the laterologs). The raw data that is measured depends very much upon the
geometry of the measurement, but these effects are corrected for by the logging company as standard.
Logging Speed. The speed at which the logs are run has an effect upon the quality of the data as a
result of statistical fluctuations in the measured radiation in the case of radiation tools, and as a
function of the measurement (sampling) interval in other tools. It can therefore be seen that logging
speed, data quality and vertical resolution are linked.
Geoff Page 2010 Page 73
Petrophysics MSc Course Notes Fluid Testing and Pressure Logs
7.1 Introduction
Formation (fluid) testing involves taking fluid samples from the formation and measuring their
pressures. It gives information on the types and properties of fluids in the formation, indicates the
presence of hydrocarbons, and provides information on the pressures of the fluids within the
formation.
There are three generic types of test, the second of which will be examined in detail as it is a Wireline, and
more recently LWD, method. The three measurement types are as follows, and are described in the order
that they are performed, in their complexity, and therefore their cost.
Perforated Permanent
Interval Packer
These testers can take fluid samples from multiple depths/sample points and keep them separated in
multiple sample chambers, up to 24 samples in a single trip. Fluids can be maintained at high pressure
both while pressure testing, sampling (by keeping the tanks at a minimum of Hydrostatic pressure) and
within the sample tanks during return to surface (by a Nitrogen gas cushion). This pressure is
maintained until the sample tank reaches a laboratory where the sample can be analysed properly. This
is important in some oils as a sudden pressure drop causes a change in the composition of the oil such
as asphaltene deposition, or phase changes.
The tools incorporate a pre-test or “drawdown” facility, where the seal between the probe and the rock
formation is tested, pressure measured and an adequate flow of fluids for sampling is checked, this
pre-test data is also used for an estimate of permeability. If the setting depth is not suitable the tool can
be reset at another depth for another try. Most tools now contain sensors that enable the fluid passing
through the tool to be analysed, distinguishing oil, water and filtrate, and allow fluid to be pumped
through the tool and exhausted to the wellbore. Only when clean formation fluid is seen will it then be
diverted to the tanks so that when a sample is taken it samples only the reservoir fluid. The tools can
cope with consolidated and unconsolidated formations, and provide very accurate fluid pressure
readings.
It should be noted that even with the newer tools, the FT is expensive to use if many fluid samples are
required. Hence, FT fluid samples are expensive.
7.3 Operation
The tool is run into the well to the depth required, which is recognized by comparing the gamma ray
readings from a gamma ray sensor attached to the tool with previously taken logs. In this way an
accurate depth may be fixed. Initially the tool takes a reading of the drilling mud pressure
(Hydrostatic). The tool is then attached securely to the wall of the borehole, by a hydraulic back-shoe
piston on one side of the tool. Opposite the back-shoe is the measurement and sampling head. This
consists of an annular seal or packer surrounding a sampling probe which contains a piston. Figure 7.4
shows details of the back-shoe and sampling head for the RCI device. A similar, arrangement applies
in the case of the LWD TesTrak in Fig. 7.5. although the size and weight of the tool (built into the
drillstring) alone is sufficient to attain a seal without the need for the back-shoe system.
The packer seals the sampling head from the drilling mud and mud-cake surrounding the tool (Fig.
7.6a). The probe containing the piston is then pressed through the mud-cake into the formation (Fig.
7.6b). The central piston is withdrawn, allowing fluids to pass from the formation into the tool (Fig.
7.6c). This fluid is then drawn into the pre-test or “drawdown” chamber, which expands reducing the
pressure in the flowline below the formation pressure. When this chamber is full pressure builds back
up until the flow line fluids are at the same pressure as the fluids in the formation, and this pressure is
then measured. Figure 7.7 shows the internal piping of the tool. Most modern tools now use a
variable volume pump for the pre-tests allowing multiple pre-tests with different volumes and
rates if required. In general it is better to test low permeability formations with lower pre-test
volumes and rates than high permeability formations.
Figure 7.4 The wireline RCI tool. Figure 7.5 The LWD TesTrak tool.
Packer
Filter
Piston
Mud-cake
Up to now only pressures have been measured. The measured pressures give an indication of the
productivity of fluids from the test depth. If it is decided that the tool should not take an operational
sample, but move on to another depth, then the pre-test chambers are emptied into the borehole, the
back-shoe is retracted, the drilling mud pressure is re-recorded, and the tool moves on to another depth. The
equalizer valve opens to allow pressures in the tool to equalize with Hydrostatic in the wellbore and prevent
“differential sticking”
If an operational sample is required, a valve is opened so that fluid flows into a sample chamber. Modern
tools may also allow the fluid composition to be measured with an array of sensors, as fluid is
pumped through the tool into the borehole. This allows a sample to be taken only after it is seen that
the formation has been emptied of invaded filtrate, resulting in a “clean” formation fluid sample. The
fluid sample is commonly between 0.5 and 1 litres, although much larger samples up to 100litres can
be taken. Once the sampling chamber is full, the valve is closed. Note that the fluids are sealed in the
sampling chamber at reservoir pressures. If another sample or more pressure data is required from
further depths, the pre-test chambers are emptied and the tool progresses. Finally, the tool is removed
from the wellbore with its sampling chambers full, and having taken a number of pressure readings at
sampled or unsampled depth points.
Geoff Page 2010 Page 78
Petrophysics MSc Course Notes Fluid Testing and Pressure Logs
The samples are usually sent to a specialist laboratory where the compositions, physical properties and
relative volumes of oil, gas, mud filtrate, and formation water can be measured.
The hydrostatic pressure is that of the drilling mud, and is recorded while the tool is at the required
depth, but has not been pressed against the wall by the packer (A). This is constant for a given depth in
the borehole, and depends upon the weight of the column of mud above it. As the mud density is
generally known, this value can be calculated and compared with the measured value.
When the probe penetrates the mud-cake, some mud is compressed between the probe and the
formation wall, leading to a transient pressure increase (B).
Fluid is drawn into pre-test chamber by the drawdown pump (C). The pressure drops because an
additional volume has been added to the system (the chamber). There may be some variation in the
pressure behaviour here as the flowing fluid is a mixture of mud-cake particles, mud filtrate, and
formation fluids of different flow characteristics. When the flow rate is equal to the rate of expansion
of the chamber a stable flowing pressure is reached (D) until the chamber is full. The lower the
permeability the lower this pressure will drop as the formation cannot deliver fluid fast enough to
keep up with the drawdown expansion. When the chamber is full (stops expanding) the measured
pressure begins to increase towards the formation pressure (E).
The pressure will then stabilize at formation pressure, which may take some time for low permeability
formations.
After the pressure measurement, the back-shoe is retracted, the equalizer valve opened, mud re-
enters the probe and the mud hydrostatic pressure is measured again (F).
Note that the pre-test chambers have a low volume (about 5-10 cm3). Hence, the fluid flowing into these
chambers is most likely mud filtrate. However, the pressure that is recorded is the true formation
pressures, as this is the pressure driving the mud filtrate into the chambers. In modern tools the volume
and flow rates can be controlled to optimize the acquisition time depending on the permeability of the
formation.
A Tight Test. If the sample is very impermeable the sampling pressure drops to near zero. In this case
it will take too long to obtain a pressure reading and the tool may stick in the borehole.
Stuck Tool. These are stationary measurements and any logging tool can get stuck if left at a given
depth for some time.
Plugging. Sand grains from the formation may enter the tool and block the flow lines, especially in
unconsolidated samples. This problem is reduced by the filter in the sampling probe, but fine grains
may still get through. Closing the tool expels fluid and flushes the probe out again.
Seal Failure. If the packer fails, the drilling mud will be sampled and the mud pressure will
continue to be recorded.
Modern FT tools which can take multiple pre-tests, and/or pump fluid through to the wellbore can get
around many of these problems. They can vary the volume and rate of the pre-test (slow and small for
tight, fast and large for permeable), and reverse pump to clean out plugging,
The sum of all four digital tracks is the same as the analogue data. A typical FT log for one depth is
given as Fig. 7.9. Note that the vertical scale is in TIME not in DEPTH.
Where a high precision quartz gauge is used the thousands are read from the analogue pressure curve in
Track 1, and the four digital tracks are then 100’s, 10’s, 1’s, and 1/10ths of a psi increments.
If we have several (>4 say) fluid pressure measurements at different depths in a fluid-bearing sand,
where the fluid is in good connection throughout the interval, we can calculate the pressure gradient in
the formation fluid. All the pressures should lie on the straight line which defines the pressure at a
given depth as a function of the depth and the density of the fluid. (Note that depth must be plotted
in true vertical depth for deviated wellbores.)
Geoff Page 2010 Page 81
Petrophysics MSc Course Notes Fluid Testing and Pressure Logs
This is applicable to all fluids (gas, oil or water) providing that the fluid in question is continuously
connected throughout the interval.
A plot of pressure on the x-axis against depth on the y-axis is used to interpret reservoir pressures as
shown in the following sections. This simple equation and the pressure versus depth plot allows us to
examine a large range of possibilities that might occur in a reservoir. It can be seen from the pressure
depth plot that the gradient of the line G = 1/ρfluid g, hence the density of the fluid represented by a line
ρfluid=1/(9.81 × G).
The six water pressures in Sand B determine the water pressure line according to Eq. (7.1), and by
fitting this equation to this pressure depth data, we can calculate the density of the water.
Note that the gradient of the lines is steeper for fluids of lower density. If one has pressure data, but
does not know the type of fluid, then one may infer it from the relative gradients on the pressure/depth
plot, or calculate the densities (gas very low, oil 0.5 to 0.9 g/cm3, water about 1 g/cm3). The
intersection of the two lines is the likely free water level, providing Sand A and Sand B are connected.
Sand B may be oil-bearing up-dip.
The data in Sand B do not form one straight line, but two. The higher is for oil and the lower for water.
The densities of each can be calculated as above. The intersection of the water and oil lines is the free
water level, as in the previous example. Note that this is about 2.5 m below the known oil water
contact OWC.
If Sand A and Sand B are in communication, the intersection of the gas line with the oil line gives the
gas oil contact (GOC). If they are not in communication, the GOC is controlled by the impermeable
barrier, and might be anywhere between the two sands.
If oil is missing in the column, then the likely gas water contact (GWC) would be where the gas and
the water lines intersect, again providing there is good communication between the two sands.
If the recorded gas pressures are lower (Fig. 7.12), the assumption that the sands are in communication
implies a GOC in Sand A, which is contrary to the log evaluation shown on the left of the figure.
Hence, the sands are not in communication.
If the recorded gas pressures are higher (Fig. 7.13), the assumption that the sands are in
communication implies a GOC in Sand B, which again is contrary to the log evaluation shown on the
left of the figure. Hence, the sands are not in communication.
8. TEMPERATURE LOGS
8.1 Introduction
The Temperature Log is a tool for measuring the borehole temperature. Temperature sensors are
attached to every tool combination that is run in a well for the measurement of the maximum
temperature (assumed to be at the bottom of the well), and many modern tools continuously measure
temperature on every run as the tool travels down the well.
Readings from a number of the maximum thermometers attached to different tool combinations and
run at different times are analyzed to give the corrected temperature at the bottom of the borehole
(bottom hole temperature, BHT).
The borehole temperature is an important parameter in the analysis of resistivity logs, but also for the
detection of fluid movement, the analysis of fluid pressures and in geochemical modelling of
formations and the maturity of hydrocarbons.
8.2 Theory
Temperature in the sub-surface increases with depth. The rate at which it does so is called the
geothermal gradient or geotherm. Typical geotherms for reservoirs are about 20 to 35oC/km (1-2 oF/1000ft),
although significantly higher values (up to 85oC/km) can be found in tectonically active areas, and lower
ones (0.05oC/km) in stable continental platforms. Hence, the bottom hole temperature (BHT) for a 3000
m well with a geotherm of 25oC and a surface temperature of 15oC is 90oC.
Note that this assumes that the geothermal gradient is constant. In practice this is rarely the case
because of differences in the thermal conductivities of rocks between the bottom of the hole and the
surface, and fluctuations in the surface temperature which penetrate the sub-surface and perturb the
sub-surface temperature. Low thermal conductivity rocks, such as shale, act as a thermal insulator and
have a large temperature gradient across them, while high thermal conductivity rocks, such as salt,
permit the conduction of heat efficiently, and have a small temperature gradient across them.
Note. Many modern sondes include a measurement of electronics temperature: This should NOT be used
as a wellbore temperature. LWD tools can also measure temperature but this is affected by the constant
flow of drilling mud through the drillstring.
The absolute accuracy of temperature measurements is low (±2.5oC), but the resolution is good
(0.025oC).
It should be noted that the best temperature logs are recorded while the sensor is being lowered into the
well on the first run. This is to reduce any temperature perturbations caused by the logging tool itself.
Temperature measurements made during drilling (MWD and LWD) consistently underestimate the
formation temperature because drilling mud is being circulated. Temperature measurements made on
wireline logs sometime after the drilling fluid circulation has stopped also underestimate the formation
temperature, but less than the MWD/LWD case as the formation is now in the process of reheating the
borehole. Measurements of temperature made by wireline logs at increasing times after fluid circulation
has stopped are closer and closer to the real formation temperature.
Various methods have been adopted in the past to correct the logged BHT to real formation
temperatures. The most common method is the Horner plot. The Horner method, plots the measured
temperature (at a given depth) from each of several logging runs, against log(T/(t+T)), where T is the
time since circulation of the drilling fluid was stopped (usually in hours), and t is the length of time of
circulation of drilling fluid prior to this. The parameter t represents the length of time that the borehole
was subjected to the cooling effects of the fluid, and T represents the time after circulation that the
borehole has had to partially reheat. On a log plot this is a straight line that intersects T/(t+T)=1 at
the formation temperature (= after infinite time T).
For example, Table 8.1 gives the drilling, logging, and temperature data for three tool combinations
that were run in a single hole one after another.
Figure 8.1 shows the Horner plot generated with this data, and the resulting formation temperature.
Run 3
Run 2
Run 1
Figure 8.1 The Horner plot for correction of BHT at 3200 m in a well using data in Table 8.1.
Use Comment
Correction of other tools The sensors of other logging tools are sensitive to temperature. The
temperature measurement can be used to correct for this, or to
recognize temperatures that are outside the operating range of the tool
and likely therefore to be erroneous.
Correction of Some parameters measured by other tools are sensitive to
measurements temperature. The best example is resistivity logs. The temperature
data is used to correct ALL resistivity data to a standard 24oC (75oF)
so they are not depth dependent and can be compared.
Hydrocarbon maturation The maturity of hydrocarbons depends upon the maximum temperature
that the organic remains have been subjected to, as well as time and
pressure.
Correlation Continuous temperature logs record differences in thermal gradient
that result from differences in the thermal conductivity of the
formations. These difference can be used for correlation.
Fluid movement Continuous logs can observe intervals of raised (or lowered)
temperature caused by the influx of hotter (or colder) fluids into the
borehole through the rock matrix, or more usually, through patent
fractures. This effect may also be due to cold drilling fluid
escaping into the rock, or injected water for pressure maintenance.
Overpressured zones Continuous logs also note the presence of overpressured zones, where
the hot overpressured fluids escape into the borehole and are noted
by a rise in the measured temperature.
9. CALIPER LOGS
9.1 Introduction
The Caliper Log is a tool for measuring the diameter (or
radii) and shape of a borehole. It uses a tool which has
1, 2, 3, 4, 6 or more extendable arms. The arms can
move in and out as the tool is withdrawn from the
borehole, and the movement is converted into an
electrical signal by a potentiometer.
The 4 arm (or dual caliper) tools are presented in a range of formats, an example of which is given in
Fig. 9.3. Note that data from the caliper pairs are shown together, and that they are different indicating
an oval hole. Also the tool rotates in the hole (the pad 1 azimuth P1AZI changes). The hole azimuth is
reasonably constant at HAZI≈1 80o, and the dip is almost vertical (DEVI is about 0o).
The ticks represent borehole volume. This information is useful to estimate the amount of drilling mud
in the borehole, and to estimate the amount of cement required to case the hole. There are engineering
approximation formulas to calculate both of these from caliper data.
Figure 9.3 Presentation of 4 arm caliper log data (After Rider, 1996).
• Contributory information for lithological assessment (see Table 9.1 and Fig. 9.4).
• Indicator of good permeability and porosity zones (reservoir rock) due to development of mudcake
in association with gamma ray log. N.B. Not always reliable as many modern drilling muds are
designed NOT to build a thick mudcake, which may form a seal differentially sticking the drillpipe.
• Calculation of mudcake thickness, hmc = (dbit – dh)/2, where h stands for the hole, in inches.
• Indication of hole quality for the assessment of the likely quality of other logs whose data quality
is degraded by holes that are out of gauge. Other log data can often be corrected for bad hole
conditions using the caliper readings, but the larger the correction, the less reliable the final data
will be. Centralized tools are designed to be about 4 inches in diameter for a standard 8.5 inch
hole, and they are designed to work with 2.25 inches of drilling mud between them and the
formation. If the hole caves to 14 inches, which is not uncommon, the distance to the formation
becomes 5 inches and the tool responses are degraded. This can be corrected for to some extent if
the caliper value is known. Tools that work by being pressed up against the side of the borehole
wall have even greater problems because the irregularity (rugosity) of the borehole wall makes it
impossible to obtain reliable readings. In both cases the recognition that a borehole has bad caving
or thick mudcake can help us judge the reliability of other tool’s readings.
• Selection of consolidated formations for wireline pressure tests, recovery of fluid samples, for
packer seating for well testing purposes, and for determining casing setting depths.
10.1 Introduction
Radioactivity is used in several different types of logging tool. There are those that measure the natural
radiation generated by the formation, such as the total and spectral gamma ray logs, and those that
measure the response of the formation to radiation generated by the tool, such as the neutron, density
and litho-density logs. This chapter will cover the total gamma ray log and the following one will
examine the spectral gamma ray log.
Each isotope of each element is given a code ZXA, where X is the elemental code. For example, carbon
has an atomic number Z=6, but 7 isotopes containing between 4 and 10 neutrons. These are 6C10 to
16 12
6C . Carbon-12, 6C , is the most common isotope, and it is stable. Most of the other isotopes are not
stable energetically, and decay to more stable elements by various processes whereby they lose energy
by expelling particles or photons. Carbon-14, 6C14, is one of these, and its decay process can be used to
date archaeological remains.
There are five main methods whereby an unstable isotope can gain stability by losing energy. These
are:
• Emission of an α particle, which is a helium nucleus 2He4, and carries two positive charges.
• Emission of a β- particle, which is a negatively charged high energy electron originating in the
nucleus together with an anti-neutrino, ν.
• Emission of a β+ + particle, which is a positively charged high energy positron originating in the
nucleus together with a neutrino, ν.
• Emission of gamma rays, γ, which are high energy photons (electro-magnetic waves) and have no
mass and carry no charge.
• Electron capture, which involves an electron being captured by the nucleus.
Under some circumstances neutrons may also be expelled from a material, but this is not a
spontaneous decay.
Gamma rays are the most important in petrophysical logging because they have the highest penetration
of all the radiations except neutrons. Their penetration ability means that they can be detected through
several centimetres of cement casing. Alpha and beta particles have very limited penetration ability,
being stopped immediately by any solid material.
Most isotopes found naturally in rocks are either stable, present in insignificant amounts, or generate
insignificant amounts of radiation. There are, however, a few which are significant. These are:
40 39
• The potassium isotope 19 K (the stable forms are 19 K and 19K41).
• The Thorium series isotopes.
• The Uranium-Radium series isotopes.
The first (K) has a single gamma emission at a single energy (1.46 MeV). The last two (TH, U) are
mixtures of unstable elements that generate each other in a series of radioactive emissions involving
gamma radiation. Hence, there is a spectrum of energies produced over the range 0 MeV to 3 MeV.
The energy spectra from the three gamma sources is shown in Fig. 10.1.
1.46 MeV
Potassium
Uranium-Radium Series
1.76 MeV
Thorium Series
2.62 MeV
0 1 2 3
Gamma Ray Energy (MeV)
The three gamma ray sources also produce different intensities of radiation per gram per second. The
U-Ra series produces 26000 photons per gram per second, the Th series produces 12000 photons/g/s,
and 19K40 produces 3 photons/g/s. One might think that the potassium would be insignificant.
However, the potassium is much more abundant than the isotopes in the other two series, which makes
its contribution to the overall radioactivity of a formation significant. Potassium is common in most
clays and some evaporites.
The most common gamma emitting lithology is shale. This is because shales are ultimately derived
from igneous rocks which have significant amounts of gamma emitting isotopes. Igneous rocks are
composed of quartz, feldspars and micas, the last two of which contain significant amounts of
potassium and occasionally also U-Ra and Th-series isotopes. The feldspars and micas alter to clay
minerals whose open lattice structure facilitates the inclusion of the larger radio-isotopes. These clays
form the principle components of shales. Hence, shales can contain up to 0.3% radio-potassium and up
to 0.01% of each of the U-Ra and Th-series isotopes. Unsurprisingly, potash beds are also highly
radioactive.
The gamma radioactivity from minerals in petrophysical logging are measured on the API scale,
described later. Table 10.1 shows typical API values for some common minerals. Note particularly
that shale and some evaporites have high gamma ray values, while sandstone and limestone have low
values.
Table 10.1 Gamma radiation from common minerals and lithologies (after Pirson, 1963).
Mineral or Lithology Composition
Gamma Radiation (API
Units)
Pure Mineral
Calcite CaCO3 0
Dolomite CaMg(CO3)2 0
Quartz SiO2 0
Lithology
Limestone - 5-10
Dolomite - 10-20
Sandstone - 10-30
Shale - 80-140+
Evaporites
Halite NaCl 0
Anhydrite CaSO4 0
Gypsum CaSO4(H2O)2 0
Sylvite KCl 500
Carnalite KCl MgCl2(H20)6 220
Langbeinite K2SO4(MgSO4)2 290
Polyhalite K2SO4MgSO4(CaSO4)2(H2O)2 200
Kainite MgSO4KCl(H2O)3 245
Others
Sulphur S 0
Lignite C H 0.849 N 0.015 O 0.221 0
Anthracite CH 0.358 N0.009 O0.022 0
Micas - 200-350
Figure 10.2 shows the range of gamma ray values generated by common lithologies. Note the
particularly high values for potash beds, which contain a large amount of potassium-40, and organic
shales, which contain enhanced uranium associated with their organic nature.
• Gamma rays with energy >3 MeV. These interact with the nucleus of the materials that they are
travelling through and are converted into an electron and a positron in the process (pair
production). The efficiency of the process is low, so these gamma rays may be measured by a
sensor. However, they contribute only small amounts to the overall signal.
• Gamma rays with energy 0.5 to 3 MeV. These gamma rays undergo compton scattering, where a
gamma ray interacts with the electrons of the atoms through which they are passing, ejecting the
electron from the atom, and losing energy in the process. A gamma ray in this range may undergo
several of these collisions reducing its energy from its initial value to an energy of less than 0.5
MeV in a stepwise fashion.
• Gamma rays with energy <0.5 MeV. These gamma rays collide with electrons of the atoms
through which they are passing, and are adsorbed. The gamma ray energy is either used to
promote the electron to a higher energy level or to eject it from the atom. This process is called
photo-electric adsorption, and is important in the Litho-Density tool.
Thus, gamma rays start with a given energy, and are either lost through pair production, or undergo
compton scattering until their energy is sufficiently low for them to be adsorbed by photo-electric
absorption.
The number of collisions, hence the reduction in gamma ray energy, and the number of gamma rays
adsorbed is directly related to the number of electrons in the materials through which the gamma rays
pass. High count rates are observed for materials with low electron densities, and low count rates for
high electron densities. The electron density is, of course, related to the mean atomic number and bulk
density of the material.
Photo-electric
Absorption
Electron
Neutron
Proton Gamma
Ray
<0.5 MeV
Gamma
Ray
>3 MeV
Gamma Pair
Ray Production
0.5- 3MeV
Gamma
Ray
Compton
Scattering
The gamma ray log measures the total natural gamma radiation emanating from a formation. This
gamma radiation originates from potassium-40 and the isotopes of the Uranium-Radium and Thorium
series. The gamma ray log is commonly given the symbol GR.
Once the gamma rays are emitted from an isotope in the formation, they progressively reduce in
energy as the result of collisions with other atoms in the rock (compton scattering). Compton
scattering occurs until the gamma ray is of such a low energy that it is completely absorbed by the
formation.
Hence, the gamma ray intensity that the log measures is a function of:
• The initial intensity of gamma ray emission, which is a property of the elemental composition of
the rock.
• The amount of compton scattering that the gamma rays encounter, which is related to the distance
between the gamma emission and the detector and the density of the intervening material.
Note that the gamma ray measurement device accepts gamma rays from almost a hemisphere that
includes the formation and the drilling mud between the formation and the sensor. Gamma rays may
therefore come from the formation at any angle from horizontal to almost vertically, and indeed may
come from the drilling mud itself (beware: some drilling muds are very radioactive!).
The gamma ray log is combinable with all tools, and is almost always used as part of every logging
combination run because of its ability to match the depths of data from each run.
11.2 Principles
The tool consists simply of a highly sensitive gamma ray detector in the form of a scintillation counter.
The scintillation counter is typically composed of a thalium activated single sodium iodide crystal backed by
a photomultiplier. When a gamma ray strikes the crystal a small flash of light is produced. This flash is
too small to be measured using conventional electronics. Instead, it is amplified by a photomultiplier,
which consists of a photocathode and a series of anodes held at progressively higher electrical
potentials, all of which are arranged serially in a high vacuum. The flash of light hits the photocathode
and causes a number of primary electrons to be produced. These few electrons still represent too small
a signal to be measured. The primary electrons are accelerated towards the first anode. For every
electron that hits the anode, a number of secondary electrons are emitted (between 4 and 8 usually).
These electrons are accelerated towards the next anode, where each of the secondary electrons produce
even more secondary electrons. This process, known as a “cascade” is repeated for each of say 10
anodes. If 6 electrons are emitted at each anode for each incident electron, we can see that a single
incident gamma ray ultimately produces 610 = 60,466,176 electrons, which represents a current that
can be amplified further by conventional amplifiers.
The whole process takes an extremely short time, but during this time the photomultiplier is saturated
and is insensitive to further gamma rays (this period is known as the dead time). However, the small
incident flux of gamma rays produced from rocks ensures that saturation is rarely if ever reached for
down-hole tools (and is sometimes corrected for at high count-rates with modern tools). Since the flash
of light and the number of primary electrons is proportional to the energy of the gamma ray, the final
current from the scintillation counter is also proportional to the energy of the incident gamma ray.
Hence, there will be a threshold gamma ray energy below which the scintillation counter is insensitive
to gamma rays.
Scintillation counters are relatively small devices, which means that the gamma ray tool can have a
relatively high vertical resolution.
11.3 Calibration
Intuitively, we might expect the units for gamma ray logging to be in gamma counts per second.
However, this results in extremely unwieldy values. Conventionally, the gamma ray log is reported in
pseudo-units called API units. The API unit is defined empirically by calibration to a reference well at
the American Petroleum Institute, University of Houston. This reference well is an artificial one that is
composed of large blocks of rock of accurately known radioactivity ranging from very low
radioactivity to very large radioactivity. The API unit is 1/200th of the difference between the highest
activity formation in the reference well, and the lowest.
Tools are run in the Houston well (test pit), and are used as standards to calibrate further tools at local
test pits. A further calibration check is also carried out at the well-site before and after the log is run,
by using a radioactive source of accurately known radioactivity a fixed distance from the tool.
It should be noted that logs from the FSU (Russia) and its satellite countries used counts per second as
its gamma ray logging units until recently, and still do in some areas. As with all FSU logs, the
interpretation and correction of these logs to API units is not straightforward.
It is common to see scales of 0 to 100 API and 0 to 150 API used in log presentations, as data greater
than 150 API is not common, and can always be handled by the use of wrap-around.
When gamma ray logging is carried out through casing and cement, a scale of 0 to 50 API is sometimes
used, as a result of the lower values measured due to the attenuation of the gamma count rate by the
casing, however beware of very high countrates caused by radioactive scale deposited by water flow in
the casing, which can generate several 100’s of API or more.
of the material. As we have seen in Chapter 10, there is a distribution of gamma ray energies, but at
distance from the emitting atom increases, the energy of the gamma rays decreases but compton
scattering until they are too low to be measured by the scintillation counter. Clearly, therefore, there is a
maximum depth of investigation for the tool that depends upon formation and mud density.
For Nuclear tool is it common to quote depths of investigation as 90% of the signal coming from within
that depth. In contrast for resistivity tools it is common to quote 50% signal depths. These conventions
are not universal, if quoted a depth of investigation it is worth asking what the number represents.
For average values of drilling mud and formation density, we can say that approximately 50% of the
gamma ray signal comes from within 15 cm (6 inches) of the borehole wall, increasing to 90% from
within 30 cm (1 foot). Hence, the depth of investigation, if defined at 90% of the signal, is around 30
cm. However, this will decrease for denser formations of the same radioactivity, and increase for
less dense formations of the same radioactivity.
Note that the zone of sensitivity is almost hemispherical, so the 30 cm depth of investigation applies
both horizontally (perpendicular to the borehole wall) and sub-vertically (sub-parallel with the
borehole wall). This has implications for the vertical resolution of the tool.
In order to smooth out some of the fluctuations a digital filter is usually applied to data over a depth
interval of data. This usually takes the form of a Gaussian (bell shape) average over 2-3 ft of data. This
slightly reduces the vertical resolution of the result. The filters can be removed if data is of sufficient
quality and the log was recorded slowly, thus enabling thinner bed responses to be seen, however care
should be taken in not miss-interpreting statistical fluctuations as thin beds.
• The size of the detector, which is quite small (about 6”-1ft, 15 – 30cm length).
• For modern computer logs the sampling frequency and digital filter applied. For older logs the
effect of a time constant that was applied to smooth the data. For conventional logging the
contribution to degradation in the vertical resolution from this is around 1 foot.
• The hemispherical zone of sensitivity of the sensor. As the sensor is sensitive to gamma rays from
a hemispherical zone, and its approximate depth of investigation of about 30 cm (1 foot) for
formations of average density, we can see that the degradation in vertical resolution from this
source will be about 2 ft.
Hence, the vertical resolution of the tool is 2-3 feet (90 cm). This does not mean that the measurements
will not detect beds thinner than 3ft, but they require this thickness to fully deflect to the true value of
the bed. In thinner beds a deflection will take place to a lessor degree depending on how thin the bed is.
This is quite a high vertical resolution for an open-hole tool, and so the gamma ray tool is good at
defining thin beds, for fine correlation, and for depth matching between logging runs. Note that
generally much finer features can be seen, however the response will not be the true value in the bed.
Note that the denser the mud used, the greater the underestimation will be, because of increased
compton scattering in the drilling mud. Barite muds are a particular problem as barite is very efficient
at absorbing gamma rays.
The measured overestimation may usually be corrected if the caliper log for the well is known. Figure
11.3 also shows the corrected gamma ray log. Comparison of the two show the degree to which the
caving has affected the gamma ray reading. Corrections are carried out using correction charts
supplied by the logging tool company. Each tool design has its own set of charts, which are drawn up
for a range of drilling fluids and tool geometries. Figure 11.4 shows an example of such a correction
chart. When run in eccentred mode the corrections are much smaller as the drilling mud contributes
less to the gamma ray signal, and has less opportunity to attenuate the gamma rays.
Figure 11.4 Gamma ray log correction chart for a 3.625 inch diameter tool with a KCl-free
drilling mud with a mud weight of ρf g/cm3 as a function of borehole diameter.
In the discussion so far, we have assumed that the drilling mud attenuates the gamma ray signal, but
does not contribute to it. While this is true of many drilling muds, it is not generally true. Potassium
chloride-based drilling muds are not uncommon. These muds have a natural gamma radioactivity
associated with potassium-40. The radiation from KCl drilling muds contributes to the total gamma
ray count rate measured, increasing it considerably. Another more recent development is formate based
muds, which may include potassium formate.
Geoff Page 2010 Page 104
Petrophysics MSc Course Notes Total Gamma Ray Log
In good quality ‘on-gauge’ boreholes the gamma ray log is consistently increased by a constant
amount. This is not a significant problem as a glance at the log header information would tell us that
KCl mud was used and we should expect slightly higher values of the gamma ray log than usual.
Figure 11.5 Effect of KCl drilling mud on the gamma ray log in an on-gauge borehole.
Problems may arise if the borehole diameter varies, leading to varying amounts of drilling-fluid
between the formation and the sensor with depth. In caved holes in radioactive formations there is
usually no effect observed as the caving effectively replaces a radioactive formation with radioactive
drilling fluid. However, a significant local increase in the gamma ray log can be observed where very
low radioactivity formations, such as some evaporites, have been washed out.
Computer programs can automatically correct for varying borehole size, mud weight and Kcl mud
content. Logs can now sometimes be corrected real time for these affects by the log acquisition system.
Always check the log header details to see if this has already been done to avoid double correction of
data.
GR (API)
0 150
Sand
Line Caving
Shale Caving
Sandy Shale
Shaly
Sandstone
Sandstone
** * * *
Heavy Mineral
Arkose SST F F or Glauconite
F Band
F F
Micaceous M M
SST M
M M
Organic/ Black
Shale Wrap-
around
Gypsum
Coal
Limestone
Dolomite
Halite
Anhydrite
Shaly
Fining-Up
SST
Shale
Sandy Line
Shale
Its main use is the discrimination of shales by their high radioactivity. Figure 11.6 shows how different
lithologies affect the total gamma ray log. Note that shales, organic rich shales and volcanic ash show
the highest gamma ray values, and halite, anhydrite, coal, clean sandstones, dolomite and limestone
have low gamma ray values. Care must be taken not to generalize these rules too much. For example a
clean sandstone may contain feldspars (arkose sandstones), micas (micaceous sandstones) or both
(greywackes), or glauconite, or heavy minerals, any of which will give the sandstone higher gamma
ray values than would be expected from a clean sandstone.
Shale volume is calculated in the following way: First the gamma ray index IGR is calculated from the
gamma ray log data using the relationship
(11.1)
Many petrophysicists then assume that Vsh =IGR. However, the value of IGR can then be entered into
the chart shown as Fig. 11.7, from which the corresponding value of Vsh may be read.
It should be noted that the calculation of shale volume is a ‘black art’ as much depends upon the
experience of the petrophysicist in defining what is the minimum (sand line) and the maximum (shale
line) values, noting that the sand line and/or shale line may be at one gamma ray value in one part of
the well and at another gamma ray value at deeper levels and which correction curve to use to convert
IGR to Vsh. When analyzing a reservoir interval, care should also be taken in using shale values that are
representative of the shale contained in the reservoir, and not e.g. values of a non-related overlying
shale.
Once the shale volume has been calculated, a threshold shale value may be defined which will divide
the well into a number of reservoir and non-reservoir zones. This zonation is combined with the zonation
that will also take place on the basis of porosity, permeability and hydrocarbon saturation.
hole logging, and allow perforating of the correct depths. A very rugged (although less accurate) GR
sonde using multiple gas tube (Geiger Muller) detectors is also sometimes run. These tools may contain
up to 12 detectors, so that if some are broken by the shock of perforating the tool will still function
sufficiently afterwards. An example of this is the PFC (Precise Formation Correlation) tool.
Note that the gamma ray readings will be less in the cased holes due to the attenuation of gamma rays by the
cement and casing. These effects can also be corrected back to “true” open hole responses by charts and
computer programs.
Potassium-40 emits gamma rays with the single energy of 1.46 MeV. This results in there being a
linear relationship between the gamma ray count rate and the content of potassium in the formation. In
potash deposits, the gamma ray reading after hole size correction gives approximately 15 API units per
1%wt. K2O.
There is no simple relationship between the gamma ray reading and the abundance of uranium in a
formation, because the energy spectrum also includes radiation from other elements in the uranium-
radium series.
12.2 Principles
It will be remembered that the amplitude of the output
from the gamma ray sensor is proportional to the energy
of the incident gamma ray. We can use this information
to measure the proportion of the total gamma radiation
coming from each of potassium-40, the uranium-radium
series, and the thorium series for a particular formation.
The spectral gamma ray tool uses the same sensor as the total gamma ray tool. The output from the
sensor is fed into a multi-channel analyzer (up to 256 channels) that calculates the amount of radiation
coming from each energy band. This is then analysed either using the older technique of 3 or 5 energy
“windows” matching the primary energies of 1.46 MeV for potassium-40, 1.76 MeV for the uranium-
radium series, and 2.62 MeV for the thorium series, or in the more modern tools by a full spectral fitting to
the shape of the spectrum from the 3 individual components over the entire energy band 0.5-3.0 MeV. These
readings represent the gamma ray radioactivity from each of the 3 sources. Their sum should be the same
as the total gamma ray value measured by the total gamma ray tool, and is coded SGR if measured with
a spectral gamma ray tool. Any combination of the three components can be summed and analyzed.
However, the most important is the sum of the potassium-40 and thorium radiation, which is called
the computed gamma ray response (CGR) or KTh
.
Figure 12.2 Measured gamma ray energy spectrum from a formation after compton scattering.
Note that there are LWD Spectral GR tools, however the thick steel of the drill collar construction
distorts the spectrum, absorbing many of the lower energy gamma rays. This can be empirically
corrected however is less accurate than wireline tools and these tools are rarely seen.
12.3 Calibration
The spectral gamma ray tool is calibrated using sources of accurately known composition containing only
K40, U238, and Th232, or a known mixture. These sources are placed next to the detector and the tool is
used to make measurements. The calibration is designed such that the calibrated readings of the tool
accurately report the amount of radiation from each of the radiation sources, and the total count rate is
calibrated to the Houston test pit values.
Track 1 is used to record the derived total gamma ray log (SGR), which is a sum of all the radiation
contributions, as well as the computed gamma ray log (CGR or KTH), which is the sum of the
potassium and thorium responses, leaving out the contribution from uranium.
Tracks 2 and 3 are used to record the calculated abundances associated with the radiation from the
individual contributions from each of K40, U238, and Th232. It should be noted that potassium is reported
as a percentage, while U238 and Th232 are reported in parts per million (ppm). Sometimes the ratios of the
elemental contributions are also plotted to help identify different minerals e.g. TH/K, U/TH and U/K.
the spectrum as lower energy gamma rays are absorbed more readily by the mud than high energy. For
this reason Spectral GR tools should always be run eccentered for the most accurate results.
The problem with KCl-based drilling mud is also apparent with the spectral tool. However, the effect
of the KCl mud will only be apparent in the SGR, CGR and potassium logs. Again, the eccentred
version is less susceptible to KCl mud.
It was seen in the discussion concerning the total gamma ray log that clean sands can sometimes
produce high gamma ray readings which would confuse them with shales. Such sandstones include
those containing feldspars, micas, glauconite, and heavy minerals including uranium-bearing ores.
The extra information supplied by the spectral gamma ray tool can, in most cases, help recognize these
situations, and calculate the amount of the particular radioactive minerals present.
Radioactive sandstones fall into one of six main groups, which are classified below, and may be
recognized using Figs. 12.4 and 12.5.
Clay-Bearing Sandstones. If clay minerals are known to be present in the rock, they may be
identified using Fig. 12.5.
Arkose sandstones. These contain feldspars, which have a significant potassium content, but a low
thorium content. The Th/K ratio will therefore be low (<1 ppm/%), as shown in Fig. 12.4.
Micaceous sandstones. These contain mica, which has a potassium composition that is less than
feldspars and a thorium content that is higher. The Th/K ratio is usually between 1.5 and 2.5 ppm/%.
Graywackes. These contain both feldspars and micas, and give Th/K ratios intermediate between 1
and 2.5 ppm/%.
Heavy mineral-bearing sandstones. The heavy minerals are often abundant in either U or Th or
both. The U and Th values are usually sufficiently high to ensure high U/K and Th/K ratios even if the
sandstones also contains potassium in the form of feldspars, micas or glauconite. Typically Th/K
values will be above 25 ppm/%, and U/K values will be above 20 ppm/%.
Figure 12.4 Thorium/potassium cross-plot for mineral identification using spectral gamma ray data.
Muscovite
& Illite
Interstratified
Muscovite & Illite
Kaolinite
& Chlorite
Weathering Bauxite
Clays Increases
1 10 100
Th/K (ppm/%)
Figure 12.5 Thorium/potassium ratio plot for mineral identification using spectral gamma ray data.
The fact that sands that do not contain clays are sometimes radioactive, as seen above, indicates that
there will be occasions when the shale volume calculated from the total gamma ray log (GR) or the
total gamma ray log from the spectral tool (SGR) will be misleading. However, we can calculate the
shale volume from the individual readings of the spectral gamma ray log (K, Th, and U), and from the
computed gamma ray log (CGR).
Equations (12.1) to (12.3) are better shale indicators than Eq. (11.1), since the random contribution of
U is eliminated. Equation (12.4), is almost never used. Note that all these equations actually calculate a
shale “Index I”, which can be corrected to true Vsh using Figure 11.7.
In pure carbonates, thorium will usually be absent because the common thorium ions are insoluble,
and potassium will also be negligible. So we can say that, if the formation has very low Th and K
abundances, the rock may be a pure carbonate. The rock, however, may contain uranium. Uranium
usually indicates material of an organic origin as organisms are extremely good at concentrating and
storing uranium. Uranium ions are either soluble or insoluble depending upon their oxidation state.
Highly oxidized uranium ions, from oxidizing environments (such as deserts) are insoluble. Thus a
carbonate, which also has low U abundance, comes from an oxidizing (maybe desert) environment.
Conversely, non-oxidized uranium ions from sub-surface depositional environments are more soluble,
and hence can be present in carbonates.
It should be noted that oxidizing environments are not conducive to the conservation of organic
material, while reducing environments not only favour the conservation of organic material, they aid
the conversion of the organic material to hydrocarbons. The source organic material for hydrocarbons
in carbonate rocks is often algal mats that are incorporated in the deposited rocks in a sub-sea
(reducing) environment and contain a significant amount of uranium.
In clay-bearing carbonate rocks high total gamma readings are not related only to the clay fraction, but
are also due to the presence of uranium-radium series isotopes of organic origin. High total gamma ray
readings are therefore not a reliable indicator of the shaliness of a carbonate. However, if the spectral
gamma ray log indicates the presence of K and Th together with the U, it may be said that the K and
Th contributions are associated with the clay content of the shaly carbonate, while the U is associated
with some organic source which was deposited in a reducing environment that favours the
conservation of organic material. Similarly, high K and Th values together with low U indicates a
shaly carbonate, deposited in an oxidizing environment which is not a favourable environment for the
conservation of organic material.
Note that K and Th must be present together for a clay to be indicated. The presence of K and no Th
(with or without U) is usually an indicator of the remains of algal mats in the carbonate, or of
glauconite. Thus, when calculating the shaliness of a carbonate, it is better to use the CGR (Eq. 12.1)
as in the case of sandstones in 12.10.1.1 above.
The actual values of K and Th in the shaly carbonate will depend upon the type of clay present, as
shown in Fig. 12.4.
Note that isolated intervals of any combination of high U, K and Th in carbonate rocks may
correspond to stylolites, which tend to concentrate uranium, organic matter, and clay minerals.
Table 12.1 shows the main interpretations of the spectral gamma ray data for carbonate rocks.
Explanation
K Th U
Low Low Low Pure carbonate, no organic matter or
oxidizing environment.
Low Low High Pure carbonate, organic matter,
reducing environment.
Low High Low Not a carbonate, or shaly carbonate
with rarer low K high Th clay
minerals, no organic matter or
oxidizing environment.
Low High High Not a carbonate or shaly carbonate
with rarer low K high Th clay
minerals, organic matter, reducing
environment.
High Low Low Glauconite carbonate, no organic
matter or oxidizing environment.
Also consider K-bearing evaporites.
High Low High Algal carbonate, or glauconite
present, organic matter, reducing
environment.
High High Low Shaly carbonate, no organic matter
or oxidizing environment.
High High High Shaly carbonate, organic matter,
reducing environment.
Note: Stylolites can locally concentrate U, clays and organic matter.
12.10.1.3 Evaporites
Large total gamma ray values are commonly associated with shales and certain types of potassium-
bearing evaporite. We can discriminate between these because the potassium-bearing evaporites have
much larger potassium abundances and zero thorium resulting from the insolubility of thorium ions in
water. Evaporites are deposited in oxidizing environments, so the uranium is usually also very low or
zero. Some K-bearing evaporites are shown in Table 12.2.
Density Pe Porosity ∆ T
Name Composition K
[DEN] [PEF] [CNL] [AC/Sonic]
(wt%)
(g/cm3) (b/e) (%) (µs/ft)
Sylvite KCl 52.44 1.86 8.51 -3 74
Langbeinite K 2 SO 4 (MgSO 4 ) 2 18.84 2.82 3.56 -2 52
Kainite MgSO4 KCl(H2O)2 15.7 2.12 3.5 >60 -
Glaserite (K Na)2SO4 24.7 2.7 - - -
Carnalite KCl MgCl2(H2O)6 14.07 1.57 4.09 >60 83
Polyhalite K2SO4 MgSO4 (CaSO4)2 13.37 2.79 4.32 25 57.5
(H2O)2
12.10.5 Diagenesis
Diagenesis is studied using the Th/K ratio derived from spectral gamma ray logs.
40
Weathering Granite
Effect Granodiorite
Diorite
30 Ultramafic
Syenite (Th=1 300,
U=2500, Th/U=0.52)
20
10
00 5 10 15 20
U (ppm)
12.10.6 Sedimentology
The spectral gamma ray log provides a large amount of data that can help discriminate between
depositional environments and allow grain size and mineralogical composition to be constrained.
Some of the more important sedimentologically relevant data are shown in Table 12.3.
The formation density log measures the bulk density of the formation. Its main use is to derive a value
for the total porosity of the formation. It’s also useful in the detection of gas-bearing formations and in
the recognition of evaporates and other minerals.
The formation density tools are induced radiation tools. They bombard the formation with radiation
and measure how much radiation returns to a sensor.
13.2 Theory
The tool consists of:
• A radioactive source. This is usually 2 curies of caesium-137 or cobalt-60, and emits gamma
rays of medium energy (in the range 0.2 – 2 MeV). The commonest tools use caesium-137 which
emits gamma rays with an energy of 0.662 MeV.
• A short spaced detector. This detector is very similar to the detectors used in the natural gamma
ray tools, and is placed approximately 7-8 inches from the source.
• A long spaced detector. This detector is identical to the short range detector, and is placed
approximately 16 inches from the source.
The gamma rays enter the formation and undergo compton scattering by interaction with the electrons
in the atoms composing the formation, as described in Section 9.3. Compton scattering reduces the
energy of the gamma rays in a step-wise manner, and scatters the gamma rays in all directions. When
the energy of the gamma rays is less than 0.5 MeV they may undergo photo-electric absorption by
interaction with the atomic electrons. The flux of gamma rays that reach each of the two detectors is
therefore attenuated by the formation, and the amount of attenuation is dependent upon the density of
electrons in the formation.
• A formation with a high bulk density, has a high number density of electrons. It attenuates the
gamma rays significantly, and hence a low gamma ray count rate is recorded at the sensors.
• A formation with a low bulk density, has a low number density of electrons. It attenuates the
gamma rays less than a high density formation, and hence a higher gamma ray count rate is
recorded at the sensors.
The density of electrons in a formation is described by a parameter called the electron number density,
ne, For a pure substance, number density is directly related to bulk density, and we can derive the
relationship in the following way.
• The number of atoms in one mole of a material is defined as equal to Avagadro’s number N
(N≈6.02× 1023).
• The number of electrons in a mole of a material is therefore equal to NZ, where Z is the atomic
number (i.e., the number of protons, and therefore electrons per atom).
• Since the atomic mass number A is the weight of one mole of a substance, the number of electrons
per gram is equal to NZ/A.
• However, we want the number of electrons per unit volume, and we can obtain this from the
number of electrons per gram by multiplying by the bulk density of the substance, ρb Hence, the
electron number density is
(13.1)
Thus, the gamma count rate depends upon the electron number density, which is related to the bulk
density of a substance by Eq. (13.1). The bulk density of a rock depends upon the solid minerals of
which it is composed, its porosity, and the density of the fluids filling that porosity. Hence, the
formation density tool is useful in the determination of porosity, the detection of low density fluids
(gasses) in the pores, and as an aid in lithological identification.
13.3 Operation
A radiation emitter and one detector are all that is necessary for a simple measurement. The early tools
had only one detector, which was pressed against the borehole wall by a spring-loaded arm.
Unfortunately, this type of tool was extremely inaccurate because it was unable to compensate for
mudcake of varying thicknesses and densities through which the gamma rays have to pass if a
measurement of the true formation is to be achieved.
All more recent tools have (at least) two detectors to help compensate for the mudcake problem. The method
of compensation is described in subsequent sections. These two detector tools were called
compensated density logs. These tools in turn have now almost completely been replaced by the Litho-
density tools (LDL, ZDL) described in Chapter 14, however the basic principles remain as described
here.
Compensated formation density tools have one focused (collimated) radiation source, one short
spacing detector at 7-8 inches from the source, and one long spacing detector 16 inches from the source
(Fig. 13.1). The source and both detectors are heavily shielded (collimated) to ensure that the radiation
only goes into the mudcake and formation, and that detected gamma rays only come from the mudcake
or formation. The leading edge of the shield is fashioned into a plough which removes part of the
mudcake as the tool is pulled up the well. The tool is pressed against one side of the borehole using a
servo-operated arm with a force of several hundred pounds force. Under this pressure and the pulling power
of the wireline winch the plough can make a deep impression in the mudcake. The heavy shielding also
doubles as a skid and a wear plate that protects the source and detectors and can be replaced easily and
cheaply when worn down.
Some tools have the source at the top, but the principle and operation is exactly the same.
The operation of this tool depends upon the detection of gamma rays that have been supplied to the
formation by the source on the tool and have undergone scattering. The natural gamma rays from the
formation add a very small contribution to this, but essentially have no effect on the measurement as the
tool source is many magnitudes stronger than any natural source.
Note in Fig. 13.1 that for the short spacing detector 90% of its signal comes from within 5 cm of the
borehole wall, which includes the mudcake signal. About 90% of the long spacing signal comes from
within 10 cm of the borehole wall. Therefore the tool has a shallow depth of investigation. We wish
to remove the mudcake signal from the measurement, and the procedure to do this is described in
Section 13.4. When the mudcake compensation has been carried out, one can see from Fig. 13.1 that
the depth of investigation has been improved, and less signal comes from the first 5 cm region.
The compensation scheme described below effectively takes the effect of the near-borehole region
(short spaced detector) away from the formation plus near-borehole region (long spacing detector).
This sounds easy but is not straightforward.
For any given borehole and tool geometry there are eight independent parameters that affect the
reading of the tool. These are:
The influence of changing Zb Ab and Amc are small, and so these parameters can be considered to be
constant. Unfortunately we cannot assume that Zmc is constant as the drilling mud (and hence
mudcake) may or may not contain barite in very significant quantities.
The number of variable parameters is reduced to three by the incorporation of ρ mc and Zmc into a single
parameter, the effective mudcake density ρ mc * , which combines the influences of both ρ mc and
Z mc . The remaining three parameters are ρ b ρ mc * and hmc .
These three parameters are represented as a function of the readings from the short and long spacing
detectors by a single and rather clever graph. The y-axis (ordinate) of the graph is the reading from the
long spacing detector, and the x-axis (abscissa) of the graph is the reading from the short spacing
detector. Points plot on this graph in different ways depending upon the various values of ρb, ρmc* and
hmc.
Imagine that there is no mudcake. The short spaced detector will make a reading of gamma count rate
that is dependent upon the short path between the source and the detector through the formation. The
long spaced detector will also make a reading, which will be lower because of the greater distance of
this detector from the source.
The readings from the long and short spaced detectors will fall on a straight line (Fig. 13.2). The
higher the readings, the less dense the formation and vice versa. So we can calibrate the straight line to
be read in terms of the bulk density of the formation from either detector. This line is called the spine of
the plot.
Now imagine that a mud of variable thickness and ρ mc * = 1.5 g/cm3 is present on the borehole wall,
and the formation density is 2.5 g/cm3. If the mudcake does not exist the detector values plot at point A
in Fig. 13.3 as expected. If the mudcake thickness increases, the short spaced detector sees more of
the less dense mudcake, and so gives a higher count rate. The long spaced detector also give a
higher count rate due to the lower density mudcake that is present, but the increase is initially
marginal because the long spaced detector is sensitive to a region which the mudcake only forms a
small part. The result is the curve shown in Fig. 13.3. Note that this curve leaves the spine at 2.5 g/cm3,
when h mc =0, forms an arc to the right of the spine, and rejoins the spine at 1.5 g/cm3 as hmc→∞. The
rejoining represents the state where the thickness of the mudcake is greater than the depth of
investigation of the tool, so that both the long and short spaced detectors read only mud. The curved
part of the graph therefore represents the complete range of mudcake thicknesses.
If the effective mudcake density remains at 1.5 g/cm3, but the formation density changes, the bottom
point where the curved part of the line joins the spine slides up and down the spine, as shown in Fig.
13.4 for formation densities of 2.0, 2.5 and 3.0 g/cm3. Each of the curved lines represents a different
formation density, leaving the spine at the formation density value for hmc=0, and rejoining the spine
together at the effective mudcake density of 1.5 g/cm3.
Figure 13.2 Formation density tool mudcake Figure 13.3 Formation density tool mudcake
correction plot (spine and ribs) with no mudake correction plot (spine and ribs) with ρb=2.5 g/cm3
present. ρmc*=1.5 g/cm3, and variable thickness mudake.
Figure 13.4 Formation density tool mudcake Figure 13.5 Formation density tool mudcake
correction plot (spine and ribs) with ρb=2.0, 2.5 correction plot (spine and ribs) with ρb=2.5
and 3.0 g/cm3, ρmc*=1.5 g/cm3, and variable g/cm3, ρmc*=1.5, 2.0 and 3.0 g/cm3, and variable
thickness mudake. thickness mudake.
If the formation density remains fixed at 2.5 g/cm3 the point where the curved portion of the line
leaves the spine is fixed at 2.5 g/cm3. Now, if the effective mudcake density changes, the point where
the curved part of the line rejoins the spine slides up and down the spine, as shown in Fig. 13.5 for
effective mudcake densities of 1.5, 2.0 and 3.0 g/cm3. Each of the curved lines represents a different
effective mudcake density, leaving the spine at the formation density value of 2.5 g/cm3 for hmc=0, and
rejoining the spine together at the relevant effective mudcake density, which may be above or below
the formation density on the spine.
Note that if the effective mudcake density is the same as the formation density, the curved portion of
the line does not exist, and the tool interprets the mudcake as formation.
In practice, only portions of the curved lines are relevant to wellbore conditions. Over the practical
range, the combination of the curves and the spine is reminiscent of a rib-cage. This type of plot is
therefore called a spine and ribs plot.
Figure 13.6 shows a spine and ribs plot for the formation density tool during a well controlled
laboratory study of the tool. Note that the data on each rib is practically independent of effective
mudcake density and composition, and that as mudcake becomes thicker, the data points lie further up
or down the rib from the spine as expected. For each design of density tool (Wireline, LWD, and
different contractors) these plots are slightly different in shape, but the basic principles are all the same.
Figure 13.6 Formation density tool mudcake correction plot (spine and ribs) from laboratory derived
calibration data.
The spine and ribs plot therefore has the advantages that (i) the variation in effective mudcake density
and mudcake thickness can be described by a single parameter (the distance up a given rib), (ii) for
each formation density, all ribs leaving one side of the graph coincide on the spine at a single point
(the effective mudcake density), and (iii) knowing the spine and ribs plot for a given tool, the long and
short detector readings can be used to obtain the corrected value of formation density without explicit
measurement of either the effective mudcake density or the mudcake thickness. The correction
procedure is automated in modern tools.
13.5 Calibration
The direct relationship between electron number density and bulk density indicates that no calibration
is necessary. However, hydrogen is anomalous, and its presence perturbs the direct relationship. For
this reason the formation density tools are calibrated. The primary calibrations are made by inserting
the tool into a block of pure limestone saturated with fresh water of accurately known density.
Secondary (check) calibrations are made in the wireline tool workshop by inserting the tool into large
blocks of aluminium, and magnesium of known density. Sleeves of varying thickness are put around the
tool to simulate mudcake thickness to check the operation of the automatic mudcake compensation
process. A portable radiation source is used at the wellsite to check the responses of the two detectors
before and after the tool has been run. Calibrations are also required for each source/tool combinations
because the half life of Cs137 is only 30 years. This means that the gamma ray output of the sources
reduce measurably over a few years and they are all slightly different.
The relationship between electron number density and the bulk density is given by Eq. (13.1). Note
that there is a linear relationship between the electron number density and the bulk density, and the
remaining parameters in the equation are either constant for a given element (A and Z), or universally
constant (N). Table 13.1 shows the values of A, Z, Z/A and 2Z/A for common elements in the crust.
Table 13.1 Atomic number and mass data for common elements in the crust.
Note that Z/A is close to 0.5 for all elements except hydrogen, which is almost unity.
We define a new parameter called the effective number density ρe, where ρe = 2 ne/N. Hence, Eq.
(13.1) becomes
(13.2)
For rocks that are composed of more than one element, Eq. (13.2) is valid providing that Z is the mean
atomic number and A is the mean atomic mass.
For most elements, the constant term is almost unity giving a one to one calibration. For hydrogen the
equation breaks down. We can examine the effect of the hydrogen on Eq. (13.2) by carrying out a
calibration of the tool in limestone saturated with fresh water, where its density is accurately known.
This calibration provides the relationship
where ρa is the apparent density (that read by the tool). For water-filled sandstones, limestones and
dolomites the tool reading is almost identical to the actual bulk density. Table 13.2 compares the
apparent density read by the tool against the actual bulk density.
Table 13.2 Comparison of apparent density measured by the formation density tool with the actual
bulk density for common mineralogies and fluids.
The density in some formations such as anhydrite, sylvite and halite should be corrected to give true
density, (but rarely are as they do not contain hydrocarbons), as do readings in gas-bearing formations
(also rarely done due to lack of accurate gas properties). Figure 13.7 shows these corrections.
Corrections also need to be made for borehole size and mud weight, which can be (and usually is)
done real-time by the acquisition systems. The correction curve for one tool is given in Fig. 13.8. (NB
All tools are different and require specific charts).
The automatic compensation (correction) for mudcake is often shown in either track 2 or track 3 on a
linear scale. This curve is included as a quality control curve, and labeled either DRHO or ZCOR. As
the value of the correction curve becomes larger, the data in the main curve will be less reliable. For the
older FDC a value greater than ±0.15 g/cm3 is considered a cutoff for reliability, this is due to the shape
of the ribs curving sharply at this point. More modern tools have a more linear rib response and
corrections are more accurate.
As the formation density tool is a log that is commonly used to calculate porosity, a pseudo-porosity
that has been calculated from the density data may be shown together with the bulk density curve and
correction curve. This is a pseudo-porosity curve because it will assume a matrix and fluid density to
be constant throughout the logged interval. Matrix densities of 2.65 g/cm3 (for quartz) and fluid
densities of 1.0 g/cm3 (for water) are often used.
A wireline tension curve is often also included in the record, because the formation density tool is
prone to stick/pull as a result of its ploughing in sticky mudcake.
The formation density tool is most often run in combination with (i) a gamma ray log, for depth
matching, (ii) a caliper log, for borehole quality control (from the density backup arm = 2arm type
pad+arm), and (iii) a neutron log, because the interpretation of the formation density tool together
with the results from the neutron log provide one of the two best lithological assessment techniques for
a reservoir.
The shallow depth of investigation of the tool makes it sensitive to borehole quality, and it is therefore
necessary to interpret the formation density log together with the caliper log to ensure that the
measured values is not an artifact of bad hole quality.
The shallow depth of investigation also implies that in porous and permeable formations, where its
main use lies, it only measures the invaded zone. This should be taken into consideration when
deciding on a fluid density (mud filtrate density) to use for porosity calculations. The fact that the tool
only measures the invaded zone in porous formations makes the tool little use for distinguishing
between formation oil and formation water. However, gas may still be detected because (i) the greater
difference in density between gas and oil or water, and (ii) the fact that mud filtrate invasion into gas-
bearing zones is never complete, and always leaves a significant amount of gas behind in the invaded
zone. LWD tools may see past shallower invasion due to the short time after drilling that the
measurement is made.
The high vertical resolution means that the log is useful for defining formation boundaries.
The bulk density ρb of a formation can be written as a linear contribution of the density of the rock
matrix ρma and the fluid density ρf , with each present is proportions (1-ф) and ф, respectively :
(13.5)
The fluid densities are usually available from FT sampling, but values of 1.0 g/cm3 for water are often
used. Remember that the tool measures the invaded zone, so the relevant fluid is the mud filtrate in most
circumstances. If available, the fluid densities should be corrected to borehole temperature conditions. If
the formation is hydrocarbon-bearing the fluid density can be calculated by
(13.6)
The value of the matrix density taken depends upon the lithology of the interval under question. For
sandstones, the density of quartz is 2.65 g/cm3, and for limestones, the density of calcite is 2.71 g/cm3.
Clay minerals have varied grain densities. Often core data is used to provide accurate matrix densities
for particular intervals.
Table 13.3 Grain (matrix) densities of some common rock forming minerals.
Care must be taken within some lithological intervals because the composition of the matrix may
change. For example, the grain density for a clean sandstone is that of quartz (2.65 g/cm3). However, if
there is a variable amount of biotite present mixed in with the sand, the bulk density of the rock can
rise to 2.84 g/cm3 because biotite has a density of 2.9 g/cm3. This scenario is encountered in some
North Sea reservoirs. Analysis of Eq. (13.3) shows that an error of 0.01 g/cm3 in the matrix density
gives an error of 0.5% in the calculated porosity, which gives an error of almost 10% if a matrix
density of 2.65 g/cm3 is assumed for a sand that has an actual density of 2.84 g/cm3 due to additional
biotite.
The Effect of Fluid Density. The porosity may also be in error if the fluid density is misjudged. The
fluid existing in the zone of the rock measured by the formation density tool is usually mud filtrate.
The density of these fresh and salt waters is approximately 1.0 g/cm3 and 1.1 g/cm3, respectively.
However, these vary with temperature and composition, so accurate values for the actual reservoir
formation water at the relevant reservoir temperatures should be used wherever possible. Such data
can be obtained from samples of reservoir fluid from FT analysis, or from the analysis of mud
filtrate, bearing in mind that the tool measures the invaded zone, so the relevant fluid density to use in
the porosity calculations is most often the mud filtrate density. Mud filtrate densities are now corrected
automatically for temperature and pressure in most petrophysical software. However, the correction
used to be carried out with the aid of a nomogram (chart).
The Effect of Gas. If gas is present in the formation, porosities can be overestimated. The density of
gasses is very low (approximately 0.0001 g/cm3 at STP) compared to aqueous fluids. If the formation is
gas-bearing a significant amount of gas is always left in the invaded zone. This gas will reduce the
mean fluid density of the invaded zone, and will cause overestimations of the porosity if a fluid
density of 1.0 or 1.1 g/cm3 is used as shown in Fig. 13.10.
This corrected value of density may then be used in Eq. (13.5) to obtain a porosity which takes
account of the shaliness of the sandstone (or limestone). The corrected bulk density of the shaly
lithology is given by
(13.7)
Calibration with Core Porosities. The most accurate porosity determinations are obtained from
laboratory measurements on cores. If there is a database of core porosities for a given well, it is often
advantageous to plot the core porosity against the formation density log derived porosity or (better) the
bulk density from the formation density log. The first type of plot should show the points lying on a
1:1 straight line. The second type of plot also gives a straight line with a negative gradient (Fig.
13.11). Non-linearity in the second type of plot may indicate problems with the formation density
log reading a varying mix of fluids and minerals, while the first type of plot can be used to indicate
that there is a better matrix or fluid density to use to transform the measured bulk density data to reliable
porosity values.
Figure 13.11 Calibration of the formation density tool against core porosities.
ranges that overlap each other and that of shales (Fig. 13.12). However, when used with the neutron
log, the combination is a very good lithological determination method (as described in Chapter 15).
If, within a given shale interval, there is a sudden change of density, the most likely explanation is that
the formations above and below the change have been deposited in a completely different
environment. The change is therefore an indication of a possible unconformity (Fig. 13.14).
13.12.6 Overpressure
Fluid overpressure works against any compaction
trend caused by the overburden pressure.
Hence, it is likely that overpressured zones will
retain a greater porosity than normally pressured
zones. If a normal compaction or no
compaction is observed in a shale over some
depth interval, and below it the bulk density
begins to decrease (or the derived porosity
begins to increase) without change in lithology,
it is likely that one has entered a zone of
overpressured fluids. In this zone the
overpressured fluids keep the porosity open,
stopping any compaction trend and reversing
it (Fig. 13.13).
Mineralogies and lithologies that lower the density locally if present in thin bands include lignite, coal,
anthracite or increased organic matter in a shale.
Mineralogies and lithologies that increase the density locally if present in thin bands include pyrite,
siderite, basalt and gneiss.
The acoustic/sonic tool can also be used to measure the porosity of the formation. However, this tool
is not sensitive to fracture porosity.
Hence, the difference between the porosities derived from these two measurements can be used as an
indicator of the extent of fracturing in a reservoir interval.
Most density tools used today also make a secondary measurement of the Photoelectic absorption of the
formation, and are of the litho-density type such as Baker Hughes Z-Density (ZDL), or Schlumberger’s
Litho-Density Tool (LDT). These tools have a caesium-137 source emitting gamma rays at 0.662 MeV,
a short-spaced and a long-spaced detector in the same way as the basic formation density tool. However,
the detectors are more efficient, and have the ability to recognize and to count separately gamma rays
which have high energies (hard gamma rays: 0.25 to 0.662 MeV) and gamma rays which have low
energies (soft gamma rays: 0.04 to 0.0 MeV).
The hard gamma rays are those that are undergoing compton scattering. The count rates of these
gamma rays (in the energy window 0.25 to 0.662 MeV) are used in the conventional way to measure
the formation density (Chapter 13). The final density value obtained is more accurate than the basic
formation density tool because the harder gamma rays are less prone to attenuation by borehole
effects, and the count rates are higher due to better detectors, most tools now use crystal detectors
similar to the GR.
The soft gamma rays are those that are undergoing photo-electric absorption. This effect can be used
to provide a parameter which is dependent upon the atomic number of the formation, and therefore
immensely useful in lithological recognition.
14.2 Theory
14.2.1 Compton Scattering and Photo-Electric Absorption
Figure 14.1 shows the energy spectra of gamma rays as they are emitted from the source, and after
travelling through various distances of the rock. At the radiation source, all gamma rays have a well
defined energy of 0.662 MeV, represented by the sharp peak in Fig. 14.1. After travelling through the
rock the gamma rays undergo compton scattering and loose energy, so the initially sharp peak moves
to lower energies. Each of the gamma rays undergoes a different number of collisions dependent upon
chance, and hence looses a different amount of energy. Thus the peak is not only moved to lower
energies, but is also dispersed (becomes wider). In Fig. 14.1, Curve A represents the initial energy
spectrum, Curve B represents the energy spectrum after the gamma rays have traveled a small distance
through the rock, and Curve C represents the energy spectrum of the gamma rays after they have
traveled an additional small distance through the rock.
However, some of the most scattered gamma rays now have energies close to 0.2 MeV. Once below
the 0.2 MeV threshold, the gamma ray can be completely absorbed by the atoms in the rock, and there
is a given probability that this will occur depending upon whether the soft gamma ray encounters an
electron under the correct conditions. This is called photo-electric absorption, and is a completely
different process from compton scattering. The result is that gamma rays attaining energies less than
0.2 MeV are eaten away from the energy distribution. Curve D in Fig. 14.1 shows the energies of the
gamma rays after they have traveled another increment of distance through the rock, where some
gamma rays have attained sufficiently low energies that they have been photo-absorbed. Curve E
finally represents the energy distributions of gamma rays as they are detected by the tool. Note only
the high energy tail of the distribution remains.
B
C
D
E
Soft Gamma
Ray Window Hard Gamma Ray Window
Low U
Low Z
Medium U
Medium Z
High U
High Z
0.662 MeV
Source
Different materials have different abilities to photo-absorb gamma rays. Figure 14.2 shows the energy
spectra of gamma rays after traveling to the detector through three media with low, medium and high
photo-electric absorption indices.
We can use the count rates measured in the higher energy shaded window (for hard gamma rays) as a
measure of compton scattering and hence the electron density of the material through which the
gamma rays have passed. This gives information about the density of the formation in the same way as
for the conventional formation density tool.
The count rate of the soft gamma rays in the lower energy shaded window in Fig. 14.2 is a measure of
the density of the material (how many electrons there are available to take part in photo-absorption)
and the rate of photo-absorption per electron. Hence, the number of gamma rays reaching the detector
in the lower energy window depends upon the effective electron density of the rock as well as the
photo-electric capture cross-section of the material.
(14.1)
The specific photo-electric absorption index Pe of a material describes the likelihood that a gamma ray
will be photo-electrically absorbed per electron of the atoms that compose the material.
Both σe. and K depend upon energy, but since they do so in the same way, their energy dependencies
cancel out, and Pe is therefore independent of energy. The photo-electric absorption index can be
approximated by the empirical relationship
(14.2)
(14.3)
The higher Pe, the higher the mean atomic number of the formation. If there are isolated Pe peaks, this
may indicate local deposits of heavy minerals especially those containing iron, or radioactive placer
deposits (uranium and thorium). If there is a general high value for the Pe curve, this may indicate the
presence of igneous or metamorphic rocks.
The lower Pe, the lower the mean atomic number of the formation. Thin bands of low Pe may indicate
coal.
Formation Fluids. The porosity and fluid saturations of rocks vary, and so it would be expected that
the measured Pe and U values might also vary with changing porosity and fluid saturation in the same
way that many other logs do. However, the values of Pe and U for the fluids commonly found in rock
are so low compared with the values for the matrix, that their influence is negligible. The one
exception to this is perhaps highly saturated brines, which may have a significant Pe value. Figure 14.3
shows how little influence up to 35% porosity changes have on the Pe values for quartz, limestone and
dolomite.
The PEF log is therefore sensitive to differences in the mean atomic number of a formation without
being sensitive to changes in the porosity and fluid saturation of that lithology. This combination
makes the PEF log an extremely good indicator of lithology.
The ability of the PEF log to accurately indicate lithology is not impaired in gas-bearing zones, where
the combination of the formation density and neutron logs may have difficulty distinguishing between
lithologies.
This log is commonly run combined with the neutron log and a gamma ray log. The combination of
the formation density log with the neutron log and the PE log is very powerful in lithological
assessment. The gamma ray log is run to aid depth matching with other logs. Figure 14.4 shows the
typical layout of this tool combination.
The PE measurement has a slightly better vertical resolution still. This is because the PE log relies almost
solely on the long spacing detector rather than both detectors. It has, therefore, a smaller ‘footprint’.
As with the formation density log, it is possible to enhance the vertical resolution of this log by slowing
down the logging speed and using modern digital data processing.
The photoelectric log is one of the two most useful approaches to lithological determination downhole.
This is because the tool is simply sensitive only to the mean atomic number of the formation, and at
the same time is insensitive to changes in porosity and fluid saturation in the rock. Hence, the absolute
Pe value may often be used to indicate directly the presence of a given lithology. This lithology may
then be checked against the other tool readings for consistency. Figure 14.5 shows the response of the
tool to common lithologies (also see Fig. 14.4).
Figure 14.5 Measurements of the photo-electric absorption (PE(F)) log for common
lithologies.
Note particularly that there are two common lithologies that cannot be distinguished by the gamma ray
log easily. These are clean sandstones and clean limestones. It is clear from Figs. 14.4 and 14.5 that
the PE log can distinguish clearly and unambiguously between clean sandstones and clean
limestones.
Also, it can be seen from Fig. 14.4 that limestone and anhydrite cannot be distinguished by the
PE(F) log. However, anhydrite will always show up with little or no porosity on the neutron log, and
will have a bulk density always above 2.9 g/cm3, which compares with a maximum of 2.71 g/cm3
possible with a clean limestone.
If there is a mixture of two mineralogies, for example a sandy limestone, a crossplot technique or a
simple mixing rule can be applied to calculate the relative proportions of the two mineralogies.
In the crossplot technique, the Pe value is plotted on the x-axis against bulk density on the y-axis
(Figure 14.6). The point lies between lines that indicate the relative proportions of three lithologies
(sandstone, dolomite and limestone).
Figure 14.6 Matrix identification in a multi-mineral system using the PE and density log data for a
formation saturated with a fluid of density equal to 1.0 g/cm3.
In the mixing rule method the relative proportions of two mineralogies may be calculated if one can
obtain Pe values for each of the two pure mineralogies from fore-knowledge, standard values, or from
log zones where each of the pure mineralogies are present. The mixing rule is simply
(14.4)
When three minerals are present (plus porosity), the Pe log values are used together with the formation
density log values. These two logs are multiplied to give the volumetric photo-electric absorption
index U (according to Eq. (14.3)). The reason for this lies in the sensitivities of each of the parameters
in Eq. (14.3). The Pe value is mass related and, as has been discussed, is sensitive to lithology but
rather insensitive to porosities and fluid saturations. By comparison, the bulk density is sensitive to
changes in porosity and fluid saturation. The product of the Pe values and the bulk density values gives
a U value, which is sensitive to both the lithological composition of the formation, as well as any
changes in porosity that are lithologically controlled.
The minerals are obtained from a cross-plot of the apparent bulk density ρmaa of the matrix against the
apparent value of U for the matrix Umaa. The apparent bulk density ρmaa of the matrix is usually obtain
from a density/neutron cross-plot (see Chapter 15). The matrix Umaa is obtained from
(14.5)
The technique is very useful in complex carbonate-sand-evaporite systems where the porosity is
controlled by the lithology, as is especially useful in gas-bearing zones, which do not alter
significantly the Pe values. The technique is less useful in sand-shale sequences.
There are many heavy minerals which can be identified in small amounts on the PE and U logs. The most
common are, siderite, haematite, pyrite, glauconite and even biotite and muscovite micas.
These heavy mineral bands are not esoteric lithologies from the point of view of the reservoir
petrophysicist, as heavy mineral bands with characteristic signatures can be very useful in the
correlation of formations between wells.
14.9.3 Fractures
Most drilling fluids have very low PE values. When this information is combined with the fact that the
density tool is pad mounted and pressed against the borehole wall, one can see that most types of
drilling mud will not have a great effect on the PE measurements. The exception is barite which has a
huge PE value and will swamp all other log responses if the tool sees barite drilling mud. Some
people have used this to detect fractures. When there are fractures, the barite drilling mud will enter
them, and the logged PE value will saturate. While the hypothesis has been shown to work, it is not
used in practice, since the barite drilling mud makes all other judgments from the log invalid. In
practice PE logs are not very useful in holes drilled with barite mud, and fractures are in general
preferentially recognized by image logs.
15.2 Theory
In neutron logging there are three processes of interest: neutron emission, neutron scattering and
neutron absorption.
(15.1)
Hydrogen
0.1
Oxygen
0.01
Silicon
0.001
4.5MeV
0.1 10 1000 105 107
Neutron Energy (eV)
Figure 15.2 The fast neutron slowing efficiency of hydrogen, silicon and oxygen atoms as a function
of neutron energy for a clean sandstone, ф = 0.15.
The initially fast neutrons (>0.5 MeV) quickly lose their energy and become slower (Fig. 15.3),
passing through stages called intermediate neutrons (102 to 105 eV), epithermal neutrons (0.1 to 100
eV), and finally thermal neutrons (<0.1 eV). In solid materials containing reasonable amounts of low
atomic mass elements, this process can happen very quickly for a given neutron (of the order of micro-
seconds). However, the time taken to slow down to a given energy will vary from neutron to neutron,
depending on the chance collisions with nuclei. Thermal neutrons are so called because they have
energies which are those that a particle has as a result of it existing a room temperatures. In other
words, they only have the small energies associated with the random kinetic motion associated with
room temperatures.
Figure 15.3 The slowing of fast neutrons with time by elastic collision with formation nuclei.
When the neutrons attain epithermal or thermal energies, collisions occur much less frequently
because the neutrons are moving from nucleus to nucleus much more slowly. Within a few
microseconds of being exposed to the fast neutron source, the formation has slowed the incoming
neutrons down to epithermal and thermal levels, and a cloud of these thermal neutrons exists in the
formation surrounding the source. Collisions continue resulting in little further loss of energy and the
slow diffusion of the thermal neutrons from the zone around the detector. During this process the
neutrons are absorbed by the formation nuclei.
and
(15.3)
In neutron logging, some tools measure the epithermal neutrons, some the thermal neutrons and some
the gamma rays emitted when a neutron is absorbed.
(15.4)
Thus, for pure water (H2O), where the atomic mass of hydrogen is 1.0 and the atomic mass of oxygen
is 16.0, the partial concentration of hydrogen (CH)mass = (2×1.0)/(1×16.0 + 2×1.0) = 1/9.
It is simple to obtain the partial concentration of hydrogens per unit volume (CH)vol,by multiplying the
partial concentration of hydrogens per unit mass by the density of the material (CH)vol =
ρb×(CH)mass.
Hence, the partial concentration of hydrogens per unit volume (CH)vol = ρbwater/9 = 1/9, because the
density of pure water is 1.000 g/cm3.
Now, the Hydrogen Index of a material is defined as the partial concentration of hydrogens per unit
volume relative to water. So, if the hydrogen index of water is constrained by the definition to be
unity, and water has a partial concentration of hydrogens per unit volume of 1/9, the hydrogen index of
a material is
(15.5)
Table 15.1 shows Eq. (15.5) in practice for a few minerals and fluids commonly found in reservoirs.
Table 15.1 Hydrogen index calculations for some reservoir minerals and fluids.
Compound Formula A, n, nH ρb HI
• If the tool is in 100% water (a large tank of the stuff, say), the HI = 1.000. This is equivalent to a
rock of 100% porosity saturated with water. So we have a fixed point that HI = 1 represents ф = 1.
• If the tool is in a pure limestone rock with zero porosity, the HI = 0.000, because there are no
hydrogen atoms in the solid matrix of calcite. So we have a second fixed point where HI = 0
represents ф = 0.
• If the tool is in a pure limestone with a given porosity ф, the hydrogen index will be directly
proportional to the amount of water in the formation. Hence HI = ф for completely water saturated
limestones.
Thus the hydrogen index is a proxy measure of porosity if the solid minerals of the formation contain
no hydrogen and if the pores are completely saturated with water. But, it is the hydrogen index that
controls the count rate observed by the neutron tool. Hence we have a tool that can measure porosity in
water saturated formations where the matrix minerals contain no hydrogen.
In practice, it is not only hydrogen that affects the passage of neutrons in a formation. As we can see
from Fig. 15.2, other rock forming elements have an effect on the slowing down of neutrons in the
formation, but their effect is small compared with hydrogen, and when it comes to neutron absorption,
both chlorine and hydrogen have a part to play. The small errors that are introduced by assuming that it
is just hydrogen that affects neutrons are overcome by calibrating the tool. The tool is calibrated in pure
limestone, and hence does not give true porosities. Instead it gives equivalent porosities that would
occur if the rock was measuring limestone. This will be discussed in further detail in Section 4.
The Hydrocarbon Effect. Note in Table 15.1 that the hydrogen indices of water and oil are
approximately the same because the effect of their different compositions on the partial concentration
of hydrogen per unit volume is approximately balanced by the difference in their density.
As with the formation density tool, the zone of investigation of the neutron tool is constrained to the
flushed zone. Hence, the tool will measure the hydrogen index associated mainly with the mud filtrate,
together with whatever remains of the hydrocarbons and formation water in the formation. Imagine, for
example that we are in an oil zone in a limestone, the mud filtrate almost completely replaces the
formation fluids such that there is a saturation of mud filtrate SXO, and a residual saturation of
hydrocarbons (1- SXO). The porosity read by the neutron tool is related to the actual porosity in the
formation by
(15.5)
If the hydrocarbon is oil, this equation reduces to фN ≈ ф, because, as we have seen in Table 15.1 the
hydrogen index of water and oils is similar; HImj ≈HIhc. This is true whether the mud filtrate is oil-
based or water-based.
(15.6)
If the hydrocarbon is methane gas with a density of 0.1 g/cm3, the equation reduces to
(15.7)
If the saturation in the flushed zone SXO = 0.7, Eq. (15.7) reduces to фN = 0.77 ф. This is known as the
hydrocarbon effect in the neutron log.
The Chlorine Effect. Some types of neutron tool measure the thermal neutrons and gamma rays
produced during the capture of neutrons. There are only two elements that are found in reservoirs that
contribute significantly to neutron absorption; hydrogen and chlorine. The presence of hydrogen in the
fluids is what we want to measure, so this is not a problem.
However, if the drilling mud, mud filtrate or formation fluids contain a significant amount of dissolved
chloride ions, as is often the case, the tool will measure a lower flux of neutrons and hence
overestimate the porosity. This is called the chlorine effect, and is present in wells drilled or logged in
the presence of drilling muds containing dissolved chlorine, or in formations where the formation
waters are particularly salty.
Many modern neutron tools are designed to minimize the affects, and corrections required for, chlorine.
The Shale Effect. Shale contain clays that have a significant amount of surface absorbed (lattice
bound) water. Hence shales can contain a significant proportion of hydrogen despite being low porosity.
The apparent porosity read from the neutron tool in shale formations is therefore always
significantly higher than it really is. This is called the shale effect, and will be discussed further later in
this chapter.
• The Gamma Ray/Neutron Tool (GNT), which is rarely seen today, except in the FSU.
• The Sidewall Neutron Porosity Tool (SNP), also rarely seen.
• The Compensated Neutron Log (CNL), which is the standard log today.
Because the tool measures the thermal neutrons and the capture gamma rays, it is effected by the
process of neutron capture by chlorine. Since drilling mud, mud filtrate and formation waters may
contain significant amounts of dissolved chloride ions, the measurement can give erroneous values.
Erroneous values take the form of overestimated porosities in formations either drilled with muds
containing dissolved chloride ions, or salty formation fluids.
The detector is sensitive to epithermal neutrons. These neutrons are not yet slow enough to take part in
absorption reactions with hydrogen and chlorine. Hence, the SNP tool readings are unaffected by the
presence of chlorine in high salinity muds and formation fluids.
Formation Wire or
other tool
Far Detector
Near Detector
Source
End of tool
or other tool
Sprung-loaded
arm
The data from the GNT tool is given in API units as shown in Fig. 15.5.
The data from the SNP tool is given in equivalent limestone porosity units, with the scale running
from approximately –15% to 45%, as shown in Fig. 15.6.
Data from the CNL tool is given in equivalent limestone porosity units, with the scale running from
approximately –15% to 45%, as shown in Fig. 15.7.
15.5 Calibration
IMPORTANT The common use of limestone porosity units as the calibrating unit for these logs
means that the reading of porosity in a limestone formation will be exact, but the reading of porosity in
these units in any other formation lithology will need correction (Fig. 15.8). Be aware that this
lithology correction may be done real-time while logging, resulting in the data being presented
in either apparent sandstone or dolomite porosity – check the log heading and scales.
Figure 15.8 Correction chart for obtaining porosity values for lithologies other than limestone.
The SNP tool is calibrated at the wellsite uses a U-shaped polyethylene baffle (artificial formation)
with a two position baffle, where the geometry of the block and baffle arrangements are known to
result in readings equivalent to 11% porosity and 22% porosity.
The calibration of the CNL tool is checked at the wellsite before and after each logging run by the use
of a neutron source of accurately known activity encased in a block of plastic, (containing lots of
hydrogen to slow down the neutrons), placed a standard distance from the detectors.
Overall the depth of investigation is a little more than the density in the range of 10-20”.
The presence of gas in a formation increases the depth of investigation because, although the gas may
contain hydrocarbons, the gas has a very low density compared to liquid water and oil, and hence a
very low hydrogen index.
The presence of shales decreases the depth of investigation because even low porosity shales contain a
lot of water (i.e., hydrogen) that is bound to the surface of the clay lattice.
The depth of investigation of the CNL tool in a water saturated formation of 35% porosity is about 12
inches, and that of the SNP tool in the same formation is about 8 inches.
Thinner beds will be noticed by the tool, but the apparent porosity will not be the same as the true
porosity of the bed.
The measuring point for these tools is half way between the source and the detector for single detector
tools, and half way between the two detectors for the dual detector tool.
As with the formation density log, it is possible to enhance the vertical resolution of this log by slowing
down the logging speed and using modern digital data processing.
The SNP and CNL tools are run pressed up against the borehole wall. In this case, the roughness of the
borehole wall due to caving or wash-out can result in the detectors or source not being pressed directly
up against the borehole wall. This will cause erroneous porosity readings. This standoff effect can
require a considerable correction of 3-5% porosity for 1 inch of standoff, therefore care should be taken
to eccenter these tools as well as possible, especially in larger wellbores.
The GNT tool is sensitive to chloride-rich muds, and its results must then be corrected for the drilling
mud, mudcake and mud filtrate.
The SNP tool is not sensitive to the chloride effect and is run pressed against the borehole wall. The
mudcake tends to increase the apparent porosity because the mudcake has a high hydrogen index. This
effect can be corrected for.
The detectors of the CNL tool are sensitive to the chloride effect. However, the tool is run pressed
against the borehole wall to minimise the effect of the drilling mud on the measurement, however
corrections still need to be applied N.B. This includes correcting for the water phase chlorides in oil
based drilling mud, which can be a significant amount. Furthermore, the use of two detectors
automatically compensates for the effect of chloride-rich mudcake and mud filtrate.
The density of the mud also affects the readings, because high density muds attenuate the radiation to
a greater extent. This is usually only a problem for the GNT tool, where the effect can be compensated
for using correction charts.
The tool is sensitive to the amount of hydrogen in the formation and to a lesser extent upon other
elements. It is assumed that the contribution to the measurement by elements other than hydrogen is
negligible, and that the contribution to the measurement from hydrogen comes entirely from the fluids
fully occupying the pore space.
However, in real rocks elements other than hydrogen that exist in the rock matrix do contribute to the
measurement (e.g., chlorine in formation water, mud filtrates and in some evaporite formations), and
hydrogen is present in the matrix (e.g., bound water in shales).
The problem is partially overcome by calibrating the tool to give the porosity in limestone units. Pure
limestone saturated with fresh water is chosen because it contains no elements which contribute
significantly to the measured signal other than hydrogen.
The porosities that are read by the tool are accurate in limestone formations that contain fresh water.
The porosities that are read by the tool in other lithologies or with other fluids need to be corrected by a
chart such as that shown in Fig. 15.8.
There are three effects that are not corrected for in the log data, that need to be briefly mentioned.
The Hydrocarbon Effect. The presence of hydrocarbon liquid (oil) does not effect the tool response
as it has approximately the same hydrogen index as fresh water. Hydrocarbon gas, however, has a
much lower hydrocarbon index resulting from its low density, and its presence will give rise to
underestimations in porosity (Fig. 15.9).
The Shale Effect. Shale contain clays that have a significant amount of bound water molecules on
their surfaces. This increases the hydrogen index of the formation. Even very low porosity shales can
give erroneously high porosity readings due to the presence of these bound waters.
The Chloride Effect. Chlorine is a good absorber of neutrons, and can lead to overestimations of
porosity if present either as formation fluid or mud filtrate.
Examples
Type
Fresh and salt water Formation water
Drilling mud
Mudcake
Mud filtrate
Hydrocarbons Gasses
Oils
Oil-based drilling mud
Oil-based mud filtrate
Oil-based mudcake
Coal
Organic-rich deposits
Bound water Shales
Water of Crystallization Evaporites
Hydrated minerals Igneous and metamorphic rocks
Table 15.2 Lithological and fluid indicators for the neutron tool.
The values of apparent porosity in shales varies considerably, but is usually higher than the apparent
porosity in carbonate or sandstone rocks (i.e., 45 to 75%). This clearly high and unrealistic porosity is a
partial indicator of shale, and can become diagnostic when combined with the gamma ray log. One
may see a slight decrease in the apparent porosity in shales due to compaction, but only over large
depth intervals.
The apparent neutron limestone porosity is affected by the amount of shale and sand, or shale and
limestone in a mixture of the two. Hence we can recognize coarsening-up and fining-up sequences in
the neutron log (Fig. 15.11). While one can calculate a shale volume from the neutron log directly, it is
not recommended because of the effect of hydrocarbon gasses which may be present to disturb the log.
Figure 15.11 The neutron log response to mixtures of shales and sandstones.
Organic mater, often present in shales may cause an even higher apparent neutron limestone porosity
than the bound water in the shales alone (Fig. 15.11).
The neutron log can detect evaporites by either their waters of crystallization. The most common
example is gypsum (CaSO4.2H2O), but also include kainite, carnalite and polyhalite. The first three of these
give apparent limestone porosities of about 60%, while polyhalite give about 25%. The apparent limestone
porosities for halite, anhydrite, and sylvite are very low (approximately -3, -2, and –3 respectively, but
also tool dependent) as they contain no water. Note that there is no effect of the chlorine atom in halite
or sylvite (Fig. 15.12).
This is a hugely important technique, and together with the photoelectric log forms the best downhole
lithological identification technique available to the petrophysicist.
Both the formation density log and the neutron log give a direct measurement of TOTAL porosity.
Hence, if they are plotted on compatible scales, they should overlie each other (Fig. 15.13).
Note that the compatible scale here is Density (1.95 to 2.95 g/cm3) and Neutron (-15 to 45% limestone
porosity units). This is the most commonly used scale range, although other wider or more restricted
ones are possible.
Note also that good superimposition will ONLY occur for clean limestone formations 100% saturated
with fresh water.
• For limestone with 0% porosity, the density log reads 2.71 g/cm3 and the neutron log reads zero.
• For limestone with 100% porosity, the density log reads 1.00 g/cm3 and the neutron log reads 100.
These two points can be plotted on a graph of density (from the density log) on the y-axis against
neutron porosity (from the neutron log) on the x-axis. We can join up the two points with a straight
line, and calibrate the line for porosity (Fig. 15.14). The graph is called the density-neutron crossplot
and the line is called the clean limestone line.
The clean limestone line only works for clean limestones because of the effect of non-hydrogen
elements in the matrix on the neutron porosity values, and the different grain densities. However, we
can take account of the different densities and correct the neutron porosity units using Fig. 15.8, to
obtain lines of the cross-plot for clean sandstones and clean dolomite, the clean sandstone and clean
dolomite lines.
Figure 15.13 The density and neutron log responses for clean formations on a compatible scale (after
Rider, 1997).
One can see from the cross-plot that for density and neutron logs plotted on compatible scales, there will
be a separation of the density and the neutron logs for sandstone and dolomite, but no separation for
limestone (Fig. 15.15). The sandstone separation is called negative and the dolomite separation is in the
other direction and slightly larger, and is called positive.
The separations are caused by the relative positions of both logs. Note that these charts and the separations
are very specific to individual neutron tool designs. Ensure that you are using the correct chart.
If the shale volume decreases due to the intermixture of sandstone, the large positive separation
decreases, crosses-over and becomes eventually the small negative separation associated with pure
sandstone.
Thus, a sequence of clearly defined sand and shale formations shows switching between positive and
negative separations in the logs.
Often the beds will not be well defined, but coarsen-up or fine-up gradually. This can be seen by a
gradual switch from one separation to another. Hence the characteristic patterns (reviewed in the
chapter on the gamma ray log) can be recognized in the separations, and conclusions may be drawn
concerning the depositional environment.
The size of the separation is actually THE BEST quantitative estimator or shale volume. It is better
than the shale volume derived from the gamma ray log. We can write an equation for shale volume
based on separations
(15.8)
where Ω stands for the separation measured in 100% sandstone, 100% shale, and at the point of
interest in the log. Note that the separations may ALL be measured on the limestone porosity scale,
ALL on the density scale, or ALL using a simple ruler on the log! Each is valid, but you must keep it
consistent for the whole equation.
15.11.5 Evaporites
Remember from the section about the
formation density tool that evaporites can
often be recognized because their often pure
nature leads to well constrained density
values. Evaporites have apparent neutron
porosity values which are either very high or
very low. The combination of well
constrained density values and very high or
very low apparent neutron porosity values
is a characteristic and diagnostic
combination for evaporites (Fig. 15.17).
There are a few designs around which have a higher sensitivity to epithermal neutrons, check with the
logging contractor. These tools have a different lithological response, and generally a lower shale
response.
LWD tools are built into large drill collars which fill most of the wellbore and displace the mud
(=hydrogen replaced with steel), and generally have much smaller wellbore corrections.
Some new tools are now available which use an electronic accelerator neutron source, instead of the
chemical source, which can provide both a thermal and an epithermal neutron porosity.
When porosities are very high, there is a lot of hydrogen, and countrates are very low. This is typically
the case in shales. This means that the output from each of the two detectors varies much more
statistically at high porosities, and when the ratio SS/LS is taken and converted to a porosity via a steep
calibration algorithm, the result can be a porosity that varies considerably and appears “noisy”. This was
particularly true for the early tools which had much lower countrates.
In order to present a better looking log, an algorithm was developed to suppress this high porosity
instability. It takes the form of a modification to the calibration curve, where the curve is switched to a
lower angle linear response above around 30% porosity (Figure 15.18).
Output
Input
The effect of this is to suppress high porosity excursions above approximately 30%, but to leave all
data below 30% unchanged. (Figure 15.19), however this suppressed curve is not volumetrically
correct and should not be used for analysis, especially in shaly reservoirs (Chapter 20).
Traditionally one service company (Schlumberger) has tended to present this suppressed curve
(NPHI) on logs, and another (Baker Hughes) the true response (CNC). This has caused some
confusion when comparing logs from different companies, as the shale response looks different.
The true responses, which should be used for analysis are TNPH (Schlumberger) or CNC (Baker
Hughes), the suppressed responses are NPHI (Schlumberger) or CNCF (Baker Hughes). Other service
company data, wireline or LWD, may vary. Check with the service provider which curves are correct.
The acoustic or sonic log measures the travel time of an elastic wave through the formation. This
information can also be used to derive the velocity of elastic waves through the formation.
Its main use is to provide information to support and calibrate seismic data and to derive the porosity
of a formation.
• Provision of a record of “seismic” velocity and travel time throughout a borehole. This
information can be used to calibrate a seismic data set (i.e., tie it in to measured values of seismic
velocity).
• Provision of “seismic” data for the use in creating synthetic seismograms.
• Determination of porosity (together with the density and neutron tools).
• Stratigraphic correlation.
• Identification of lithologies.
• Facies recognition.
• Fracture identification.
• Identification of compaction.
• Identification of over-pressures.
• Identification of source rocks.
The tool works at a higher frequency than seismic waves, therefore one must be careful with the direct
comparison and application of acoustic log data with seismic data.
16.2 Theory
When the sound energy arrives at the receiver, having passed through the rock, it does so at different
times in the form of different types of wave. This is because the different types of wave travel with
different velocities in the rock or take different pathways to the receiver. Figure 16.1 shows a typical
received train of waves. The transmitter fires at t = 0. It is not shown in the figure because it is masked
from the received information by switching the receiver off for the short duration during which the
pulse is transmitted. This is done to ensure that the received information is not too complicated, and to
protect the sensitive receiver from the high amplitude pulse. After some time the first type of wave
arrives. This is the compressional or longitudinal or pressure wave (P-wave). It is usually the fastest
wave, and has a small amplitude. The next wave, usually, to arrive is the transverse or shear wave (S-
wave). This is slower than the P-wave, but usually has a higher amplitude. The shear wave cannot
propagate in fluids, as fluids do not behave elastically under shear deformation. These are the most
important two waves. After them come Rayleigh waves, Stoneley waves, and mud waves. The first two
of these waves are associated with energy moving along the borehole wall, and the last is a pressure
wave that travels through the mud in the borehole. They can be high amplitude, but usually arrive after
the main waves have arrived and are usually masked out of the data. There may also be unwanted P-
waves and S-waves that travel through the body of the tool, but these are minimized by good tool
design by (i) reducing their received amplitude by arranging damping along the tool, and (ii) delaying
their arrival until the P-wave and S-wave have arrived by ensuring that the pathway along the tool is a
long and complex one.
The data of interest is the time taken for the P-wave to travel from the transmitter to the receiver. This
is measured by circuitry that starts timing at the pulse transmission and has a threshold on the receiver.
When the first P-wave arrival appears the threshold is exceeded and the timer stops. Clearly the
threshold needs to be high enough so that random noise in the signal does not trigger the circuit, but
low enough to ensure that the P-wave arrival is accurately timed.
There are complex tools that make use of both P-waves and S-waves, and some that record the full
wave train (full waveform logs). However, for the simple acoustic log that we are interested in here,
only the first arrival of the P-wave is of interest. The time between the transmission of the pulse and
the reception of the first arrival P-wave is the one-way time between the transmitter and the receiver.
If one knows the distance between the transmitter (Tx) and the receiver (Rx), the velocity of the wave in
the formation opposite to the tool can be found.
In practice the acoustic log data is not presented as a travel time, because different tools have different
Tx-Rx spacings, so there would be an ambiguity. Nor is the data presented as a velocity. The data is
presented as a slowness or the travel time per foot traveled through the formation, which is called delta t
(∆t or ∆T), and is usually measured in µs/ft. Hence we can write a conversion equation between velocity
and slowness:
where the slowness, At is in microseconds per foot, and the velocity, V is in feet per second.
The velocity of the compressional wave depends upon the elastic properties of the rock (matrix plus
fluid), so the measured slowness varies depending upon the composition and microstructure of the
matrix, the type and distribution of the pore fluid and the porosity of the rock. The velocity of a P-wave
in a material is directly proportional to the strength of the material and inversely proportional to the
density of the material. Hence, the slowness of a P-wave in a material is inversely proportional to the
strength of the material and directly proportional to the density of the material, i.e.;
(16.2)
The strength of a material is defined by two parameters (i) the bulk modulus K, and (ii) the shear
modulus µ.
The bulk modulus, K is the extent to which a material can withstand isotropic squeezing (Fig. 16.2a).
Imagine an amount of material subjected to an isotropic pressure P1. Now let the isotropic pressure
increase to a pressure P2. The material will compress from its initial volume v1 to a new smaller
volume v2. The bulk modulus is then given by;
(16.3)
where ∆P is the change in pressure, and ∆v is the change in volume. Thus ∆P is the change in pressure
that causes ∆v change in volume.
The shear modulus, µ is the extent to which a material can withstand shearing (Fig. 16.2b). Imagine
an amount of material subjected to a isotropic pressure P1. Now apply a shear stress (non-isotropic
pressure) Ps to one side of the sample. The material will shear to the new shape, and its overall length
will increase from its initial length l1 to a new larger length l2. The bulk modulus is then given by;
where γ is the shear strain. The application of the shear stress Ps causes the development of a shear
strain γ.
Detailed analysis of the velocity and slowness of P-waves in a material shows that;
(16.5)
(16.6)
Imagine a pulse emanating from a Tx on an acoustic tool. It will travel through the drilling mud and
encounter the wall of the borehole. The P-wave travels well through the mud at a relatively slow
velocity, Vm, as the mud has a low density. The S-wave will not travel through liquid mud. At the
interface it is both reflected back into the mud and refracted into the formation. The portion of the P-
wave energy that is refracted into the formation travels at a higher velocity, Vf, because the density of
the rock is higher. This is depicted in Fig. 16.3. We can use Snell’s law to write;
(16.7)
and at the critical angle of refraction, where the refracted wave travels along the borehole wall, R= 90o,
so;
(16.8)
Hence, if the velocity of the elastic wave in the formation changes, the critical angle, i, will also
change.
The velocity of the refracted wave along the borehole wall remains Vf. Each point reached by the wave
acts as a new source retransmitting waves back into the borehole at velocity Vm.
(a) (b)
Formation Borehole Formation Borehole
Vf Vm Vf Vm
i i
R r r
R
• The time taken for elastic wave to reach Rx1: TRx1 = A+B+C
• The time taken for elastic wave to reach Rx2: TRx2 = A+B+D+E
• The acoustic interval transit time: ∆T = (TRx2 - TRx1) = A+B+D+E – (A+B+C) = D+E-C.
• If tool is axial in borehole: C = E, so ∆T = (TRx2 - TRx1) = D
The problem with this arrangement is that if the tool is tilted in the hole, or the hole size changes (Fig.
16.6), we can see that C ≠ E, and the two Rx system fails to work.
A
Tx
Rx C
Tx A
C
Rx1
D
E
Rx2
D
E
A typical pulse for the BHC is 100 µs to 200 µs, with a gap of about 50 ms, giving about 20 pulses per
second. There are four individual Tx-Rx readings needed per measurement, so 5 measurements can be
made per second. At a typical logging speed of 1500 m/h (5000 ft/h), gives one reading per 8 cm (3
inches) of borehole. Several versions of the BHC are available with different Tx-Rx distances (3 ft.
and 5 ft. being typical), and the Rx-Rx distance between pairs of receivers is usually 2 ft.
In practice to simplify tool design, and because the exact distance required between receivers is related
to the refraction angle, Rx1 and Rx4 are frequently physically the same transducer, as are Rx2 and Rx3,
resulting in a 2 x Transmitter and 2 x Receiver tool.
It was recognized that in some logging conditions a longer Tx-Rx distance could help. Hence
Schlumberger developed the long spacing sonic (LSS), which has two Tx two feet apart, and two Tx
also two feet apart but separated from the Tx by 8 feet. This tool gives two readings; a near reading
with a 8-10 ft. spacing, and a far reading with a 10-12 ft. spacing (Fig. 16.8), both of which are
uncompensated for instrument tilt. Computer systems are able to calculate a “virtual” second dataset
using depth delaying techniques as if the tool had a set of upper transmitters and apply the
compensation. This is known as Depth derived Borehole Compensation (DDBHC).
RxU
RxL
TxU
TxL
Table 16.1 The names and mnemonics of older industry acoustic tools.
Company
Tool Mnemonic
16.4 Calibration
The tools are not actually calibrated as the measurement depends only on accurate timing, and the
physical length of the tools. However the tool readings are verified inside the borehole opposite beds of
pure and known lithology, such as anhydrite (50.0 ts/ft.), salt (66.7 ts/ft.), or inside the casing (57.1
ts/ft.).
The log will also show the integrated travel time (TTI). This value is derived simultaneously with the
main measurement, and is the mean travel time in milliseconds. It is commonly shown by pips on the
right side of the depth column (in slightly different styles by different companies). The small pips occur
every 1 millisecond and the larger ones occur every 10 milliseconds. The main advantage of this is that
the average travel time between two depths can be found by simply counting the pips, which is really
useful when comparing acoustic logs to seismic sections. Likewise adding the pips together for a certain
depth interval, and dividing by the thickness of the interval allows the mean velocity of the interval to
be simply calculated. In the comparison it MUST be remembered that the pips give one-way travel
time, and must be doubled to be compared to the two way travel time given in seismic data.
The accuracy of the travel time integrator can be checked on a log in a thick homogeneous formation.
First, count the pips over the thickness of the interval to get the TTI value in milliseconds. Then read
off the eyeballed mean ∆t in the interval in microseconds per foot, multiply by the thickness of the
interval and divide by 1000. The value should agree with the TTI reading.
Modern array tools can derive ∆t in a variety of ways, with different vertical resolutions. Typically the
presented data is averaged across an array of 8 receivers at 6 inch spacing between each, resulting in a
3 1/2ft vertical resolution, however data can be processed down to the receiver spacing of 6 inches, or
2ft to match older data.
16.9.1 Noise
Noise from stray electrical fields, the electronics package or derived from mechanically generated
noise in rough holes can trigger the detection circuitry before the first arrival, causing a false (shorter)
apparent first arrival. To limit this effect, all receiver circuits are disabled for 120 microseconds after
the pulse. As the remaining time for the possibility of a noise spike occurring is greater for the far
detector than the near one, most noise spikes occur for the far detector, which leads to values of ∆t that
are too small. These are seen as single data point spikes of low ∆t in the log (Figs. 16.10 and 16.11).
16.9.2 ∆t Stretch
In heavily attenuating formations the value of ∆t can be slightly too large due to the thresholding
method employed by the detection circuitry. However, this problem is rarely significant, and is
impossible to detect from the log.
Noise, stretch, and cycle skipping are not such a problem for array tools due to a different method of
deriving the ∆t (Section 16.11).
Near Detector
Receivers Off
∆t = T2-T1
Far Detector
Receivers Off
Near Detector
Receivers Off
∆t = T2-T1
Far Detector
Receivers Off
Acoustic tools generate signals omni-directionally, and are generally run centralized so that signals from
all sides of the wellbore arrive simultaneously and enhance each other, maximizing the signal strength
available for analysis, and minimizing any noise.
If a low velocity altered zone exists, the Tx-Rx spacing must be large to ensure that the P-wave from the
virgin formation arrives before that from the altered zone. In this case an LSS should provide better
data than a BHC type log.
If a high velocity altered zone exists, there is no solution whatever older type of log is used. The
measured log value will be that for the altered zone, however array tools may see past this zone.
A graph can be drawn to show the reliability of tools affected by mud and altered zone arrivals (Fig.
16.12). It is clear that the LSS has the better performance. However, the greater spacing means that
arrival waves are weaker and therefore more prone to cycle skipping and noise spikes.
Figure 16.13 shows an example of altered zone arrivals affecting a BHC log but not an LSS log.
200
150
Reliable Area
Under Curve
100
6 8 10 12 14 16 18 20
Hole Diameter (inches)
Fig. 16.13 Example of altered zone arrivals affecting the BHC log.
It must be remembered that the acoustic log gives a one-way travel time, and the seismic technique gives
a two-way travel time.
A time-depth curve can then be obtained by taking a weighted sum of the interval velocities with
depth, which will give the total time to a given depth and plotting this against depth. Figure 6.14
shows an example of such a curve, where the acoustic interval transit time is given in parentheses next to
the depth column.
The time-depth curve can then be compared against the velocity analyses from the seismic data, or can be
used in place of velocity analyses in seismic processing.
It should be remembered that the observed seismic trace is primarily a record of the ability of
interfaces between formations to reflect elastic waves. This ability is called the reflection coefficient
R. The reflection coefficient depends upon the properties of the rock either side of the interface, and in
particular on its acoustic impedance. The acoustic impedance is the product of the seismic velocity and
the density of the rock.
Thus, if we can derive the density and seismic velocity of a set of formations from logs, we can derive
a synthetic seismogram. The procedure is as follows.
• Obtain a density log, ρ(z), for the interval of interest. This is best obtained from a density log, but
approximations can be made from the acoustic log by using the acoustic log to derive porosity, and
then if the densities of the rock matrix and formation fluids are known, the density of the rock can be
calculated.
• Convert the density log in depth to that against two-way time using the TTI information from the
acoustic log.
• Obtain the elastic wave velocity log, V(z), from the acoustic log using Eq. (16.1).
• Convert the elastic wave velocity log in depth to that against two-way time using the TTI
information from the acoustic log.
• Multiply these two values for each depth to give the acoustic impedance log, AI =ρ(TWT)V(TWT).
• Calculate the reflection coefficient for each interface from the acoustic impedance log using;
(16.9)
where the subscript 2 refers to the formation below an interface, and the subscript 1 to the
formation above it.
Note that the reflection coefficient will only exist at interfaces, and is zero in between.
• Apply a multiplying factor to the reflection coefficient log to account for the fact that the seismic
response will be attenuated with depth. This factor reduces with increasing two-way travel time
(i.e., as depth increases).
• Convolve (multiply) the modified reflection coefficient with a chosen zero phase wavelet that
represents the seismic data with which you wish to compare the synthetic seismogram.
Note that the reverse procedure is also possible, and is used to obtain information about lithology and
porosity (via density) (i.e., information about formation properties) from the seismic traces that are
sensitive to the interfaces between formations. This is called acoustic impedance inversion, and is the
domain of the geophysicist.
The velocity of elastic waves through a given lithology is a function of porosity. Wyllie proposed a
simple mixing equation to describe this behaviour and called it the time average equation (Fig. 16.16). It
can be written in terms of velocity or ∆t:
(16.10)
(16.11)
Hence;
(16.12)
where ∆t is the transit time in the formation of interest, ∆t p is that through 100% of the pore fluid, and
∆t ma is that through 100% of the rock matrix, ф is the porosity, and the velocities are analogous. A
list of input values to these equations for common lithologies and fluids is given as Table 16.2.
Table 16.2 Values for ∆t and V for use in Wyllie’s time average equation.
The Wyllie time average equation gives porosities that are overestimated in uncompacted formations
(indicated by the rule of thumb that adjacent shale beds have ∆t values greater than 100 microseconds
per foot). An empirical correction Bcp is then applied:
(16.13)
where Bcp is approximately equal to the value of ∆t in the adjacent shales divided by 100.
Fig. 16.16 The wave path through porous fluid saturated rocks.
The compaction correction factor can also be obtained from other logs:
• From the density log: A density-acoustic cross-plot in clean water-bearing formations close to
the zone of interest establishes a line that can be scaled in porosity units. But this is little better than
using the density log to calculate the porosity directly.
• From the neutron log: The compaction factor is the ratio of the porosity from the uncorrected
acoustic log to that from the neutron log in clean water-bearing formations close to the zone of
interest. Again, this is little better than using the neutron log directly to obtain the porosity.
• From the resistivity log: Obtain the porosity from the resistivity log in clean water-bearing
formations close to the zone of interest using Archie’s law and known values of its parameters.
The compaction factor is then the ratio of the porosity from the acoustic log to the porosity from the
resistivity log. Again, this is little better than using the resistivity log directly to obtain the
porosity.
Even though these methods are not so useful to obtain a corrected acoustic porosity, they are useful if one
wants to calculate the correction factor for its own sake.
(16.14)
This provides a much superior accuracy porosity over the entire range of geologically reasonable ∆t.
Figure 16.17 shows the Raymer-Hunt equation for some typical lithologies.
Fig. 16.17 The Raymer-Hunt equation for calculating porosity from transit time. Here the following
data have been used: ∆t = 53 (sandstone matrix ), 45 (limestone matrix), 40 (dolomite matrix), 186
(fluid), all in µs/ft.
Occasionally, there is good core coverage in a well, so core porosities are available. If this is the case, it
is useful to calibrate the acoustic log against the core porosity. A cross-plot of core porosity against the
transit time at the same depth should produce a straight line that can be extrapolated to the x-axis to
give a value for the local matrix transit time ∆tma (Fig. 16.18). This is best done for each obvious
lithology in the well providing there are enough core determinations to ensure that the cross-plot for
each lithology is worthwhile. Cross-plots can also be carried out between the core resistivity while
100% saturated with water, Ro, and the transit time from the acoustic log. This should also give the local
matrix transit time, and is a verification of the first plot.
The core porosity, acoustic transit time cross-plot can also be used to calibrate the porosities derived
from the acoustic log between the cored points in a given well, but should not be extrapolated out of the
cored interval or to other wells.
The acoustic log is sensitive only to the primary intergranular porosity. By contrast, the density and
neutron logs record the total porosity. The difference between the two measurements, therefore, can be
used to calculate a value for the secondary porosity, whether it be isolated vugs in carbonates or
fractures. This is called the secondary porosity index (SPI or ф2), and is defined:
(16.15)
The effect of shales is very variable. This is because is depends upon the density of the shales, which
varies a lot. Young shales are generally under-compacted and low density, tending to increase the
transit times and hence give slightly higher acoustic derived porosities. Exactly the opposite is the
case for ancient compact shales with high densities, which give lower transit times and smaller
porosities. The effect of shales on the porosity from the acoustic log is not as great as the effect of gas.
Fig. 16.20 Subtle textural and structural variations in deep sea turbidite sands shown on the acoustic
log (after Rider).
The acoustic log data is diagnostic for coals, which have very low velocities, and evaporites, which
have a constant, well recognized velocity and transit time (see Table 16.2).
It is best to use the acoustic log with other logs if lithological identification is important.
The main characteristics of the acoustic log are shown in Fig. 16.21.
16.10.7 Compaction
As a sediment becomes compacted, the velocity of elastic waves through it increases. If one plots the
interval transit time on a logarithmic scale against depth on a linear scale, a straight line relationship
emerges. This is a compaction trend.
16.10.8 Overpressure
The acoustic log can be used to detect overpressured
zones in a well. An increase in pore pressures is
shown on the acoustic log by a drop in acoustic
velocity or an increase in acoustic travel time
(Fig. 16.23).
The ∆t or slowness of the formation is extracted in a different way from array data. Instead of
attempting to determine the precise start (Travel time) of each waveform, which is subject to cycle
skipping and noise, the signals from all the receivers are compared simultaneously (Figure 16.25).
As each of the receivers is at a fixed/constant spacing from the next (usually 6 inches), the first
arrivals of each of all the waveforms line up, and the slope of the alignment is related to the
slowness to be determined (Figure 16.25 (A)), a steeper angle means a larger/slower ∆t. A
correlation window of data is compared across the array at varying slopes, and a correlogram plot
(B) then shows the highest correlation value at the angle at which all data aligns; this is the required
slowness value.
40 Slowness 240
This array correlation process, know as “semblance” or “Nth Root” also allows the slowness of
later wave arrivals to be determined, such as the Shear and Stoneley slowness’s.
This process even allows the array acoustic to be logged through casing (as long as the casing is
reasonably cemented to provide energy coupling through to the formation). An additional
correlation peak is seen for the steel of the casing at a slowness of 57µsec/ft, (Figure 16.26) which
can be excluded by a simple cutoff from being used as the formation signal, although of course this
is difficult if the formation signal also has a similar slowness.
17.1 Introduction
The whole of resistivity logging is based upon a few very important equations which are introduced in
this section. The equations, which are known as the Archie Equations, relate the resistivity of a
formation to the resistivity of the fluids saturating a formation, the porosity of the formation and the
fractional degree of saturation of each fluid present. As always, the story begins with Ohm’s Law.
Ohm’s Law states that the current flowing from point A to point B in a conductor I is proportional to
the difference in electrical potential ∆E between point A and point B. The constant of proportionality is
called the electrical conductance c. Current is measured in amperes (A), potential difference in volts
(V), and conductance in siemens (S).
(17.1)
(17.2)
(17.3)
Thus, if we take a cylindrical rock sample with two flat faces A and B, and set a potential difference
∆E =EA-EB between its end faces, a current I will flow through the rock from face A to face B (Fig.
17.1). If we measure the current and the potential difference, we can calculate the resistance of the
rock sample using Eq. (17.1).
∆E
I A B
• If the resistance is high, a given potential difference ∆E will only give a small current I.
• If the resistance is low, a given potential difference ∆E will give a high current I.
The value of resistance is a property of the material which describes how much the material resists the
passage of a current for a given applied potential difference.
• If the length of the sample is doubled, one can see that the resistance of the sample to the passage
of a current should also double.
• If the area perpendicular to the current flow doubles (the area of the end face in this example),
there is twice the material for the current to pass through, the resistance of the sample to the
passage of the current should therefore fall to a half of what is was before.
So the resistance (and therefore conductance) depend upon the size of the sample.
If we take the resistance per unit length and area, we can remove the effect of the dimensions of the
sample. The value we obtain is then only a function of the property of the material and not its
dimensions. The resistance per unit length and area is called the resistivity R, and can be expressed as
(17.4)
Note that a conductivity C can also be defined as the reciprocal of the resistivity R, and therefore
(17.5)
In petrophysical logging of electrical rock properties there are two main types of tool. One type
measures resistivity directly, and the result is given in ohm.m (Ω.m). The other type measures
conductivity directly, and the result is given in either siemens per metre (S/m), or more often in milli-
siemens per metre (mS/m). The two measurements are, of course, measuring the same property of the
rock, and can be interconverted using
(17.6)
(17.7)
The SI units mS/m are commonly labeled as mmho in oilfield units (mho = ohm backwards !!)
All have a high electrical resistivity (electrical insulators) except the formation water and water-based
mud filtrate, which are good electrical conductors and have a low electrical resistivity.
The resistivity of the reservoir rocks therefore depends only upon the water or water based mud filtrate
occupying its pore space.
There are very rarely layers of metallic conductors in rocks, such as mica-pyrite in some parts of the
Middle East. When present very small quantities of these layers can have a very large effect on the
resistivity logs as they have extremely high conductivities. Special techniques need to be used to analyse
this data, which are beyond the scope of this course.
Uninvaded Formations. For uninvaded formations, the measured bulk resistivity of the rock depends
only upon the amount of the aqueous formation fluids present in the rock, and the resistivity of those
aqueous fluids. Since the amount of formation fluids depends both on porosity ф and water saturation
Sw, we can say that the resistivity of the formation Rt depends upon porosity , water saturation Sw, and
the resistivity of the formation water Rw. This resistivity is called the true resistivity of the formation. It
is the resistivity of the formation in the uninvaded zone, where the rock contains some saturation of oil
So, gas Sg, and water Sw, and where So+Sg+Sw=1.
The aim is to use knowledge of the resistivity of the formation, together with independent knowledge
of the porosity and resistivity of the formation waters, to calculate Sw, and hence enable ourselves to
calculate the STOOIP.
Typical values of Rt range from <0.2 to 2000+ ohm.m (5000 to 0.5 mS/m).
The uninvaded zone of formations is commonly only measured directly by the most deeply penetrating
electrical logging tools. The shallower investigating tools measure the invaded zone. Hence, if one
wants a resistivity reading for use in STOOIP calculations, one should always choose the deepest
penetrating electrical tool of those that have been run.
Invaded Zones. In most cases there is an invaded zone, where the formation fluids have been
disturbed by the drilling fluid. The resistivity of the formation in this zone depends upon the resistivity
of the mud filtrate Rmf, the resistivity of any remaining formation water Rw, the saturation of the mud
filtrate SXO, the saturation of the remaining formation water Sw (if any), and the porosity of the rock ф.
If these values and the depth of invasion are known, the resistivity measured in the invaded zone can
be corrected to account for the presence of the mud filtrate.
The symbols used in electrical logging in the invaded borehole environment are shown in Fig. 17.2.
In the former case the temperature is given directly at a given depth. However, this is not used to
calculate the mud resistivity as the newer temperature logs also measure directly the mud resistivity
as it makes the temperature log.
In the latter, and more common, case the BHT allows us to calculate a mean geothermal gradient
for the borehole. The formation temperature can then either be calculated directly, or obtained with a
nomogram such as that given in Figure 17.3.
The Effect of Salt Composition. The resistivity of the formation fluids depends upon the
concentration and type of salts dissolved in it. We know the concentration and type of dissolved
solids in the formation water usually from chemical analysis of samples obtained by the FT. The
resistivity of this solution at a given temperature and pressure can be obtained by making up a
synthetic brine to the recipe indicated by the chemical analysis of the FT sample, or more often,
by the use of equations relating the composition to the fluid resistivity.
For example, a solution contains 20,000 ppm NaCl, 10,000 ppm KCl and 1000 ppm MgSO4. The
multipliers are Na(1.00), Cl(1.00), K(0.9), Mg(1.63), SO4(0.64). The total NaCl equivalent is
20,000× 1+20,000× 1+10,000×0.9+10,000× 1+1000× 1.63+1000×0.64 = 30,635 ppm NaCl.
The Variation of Formation Fluid Resistivity with Temperature. The resistivity of aqueous
formation fluids varies significantly with temperature. The resistivity decreases approximately 4% per
degree centigrade increase. Clearly, if the BHT is 200oC, the change in the measured resistivities of the
formation fluids can be as much as 800%. This will cause the resistivities measured with downhole
tools to decrease steadily down the borehole.
In order to compare fluid resistivity measurements they can be corrected to read the resistivity they
o o
would read at some constant temperature (24 C/75 F is often used). To do this we need to convert the
raw resistivity data.
Figure 17.5 shows a typical chart for this type of correction. It is based upon Hilchie’s equation for the
variation of the resistivity of aqueous fluids with temperature. Hilchie’s equation relates the resistivity
of an NaCl solution at one temperature to that at another.
(17.8)
Note: for a quick estimate “X” is often approximated as 6.77 when using ºF, and 21.5 when using ºC.
The corrections for temperature are generally included in most modern log analysis programs,
simplifying the process. For every fluid resistivity a temperature is requested, and often a temperature
gradient (vs depth – NB. All gradients/depths MUST be TVD depth) in order to continuously correct
the values over long log sections. The most important point is that:-
The Variation of Drilling Mud Resistivity with Temperature. The resistivity of the drilling mud is
measured by many tools and is available directly for calculation at a given depth. Tools that measure
the drilling mud resistivity include the newer temperature tools and some FT tools. Samples of mud are
also usually measured directly by the logging engineer at the time of the data acquisition and included
on the log headings.
The Variation of Mudcake and Mud Filtrate Resistivity with Temperature. The mud filtrate and
mudcake resistivities may be estimated from the mud density (mud weight) and the resistivity of the
drilling mud using the nomogram in Fig. 17.6. Note that this nomogram is valid for 24oC/75oF, however
the best values are the directly measured values obtained from the log headings.
Archie observed that the bulk resistivity of a rock Ro fully saturated with an aqueous fluid of resistivity
Rw is directly proportional to the resistivity of the fluid
Ro =F R w (17.9)
The constant of proportionality F is called the Formation Factor and describes the effect of the
presence of the rock matrix.
It can be immediately seen that F = 1.00 for a rock with 100% porosity, i.e., no matrix, just 100%
fluid.
If we take 100% fluid and slowly add grains of rock, the porosity decreases. However, the insulating
grains of rock have negligible conductivity (infinite resistivity) compared to the conducting fluid.
Hence, Ro will increase, which implies that F is always greater than unity in a porous medium such as
a rock. In real rocks F takes values usually between 20 and 500. Note that formation factor has no
units because it is the ratio of two resistivities.
One way of interpreting the formation factor is therefore as a factor that describes the extent to which
electrically insulating mineral grains ‘dilute’ the conducting fluid, making the bulk material more
resistive than the fluid alone. The formation factor includes both the effect of the variable porosity and
the effect of the tortuous pathways that the current is forced to take through the conducting fluid due to
the presence of the insulating rock grains. It can be seen, therefore, that the formation factor is related to
the porosity of the rock and the connectivity of the pore spaces. The natural complexity of pore systems
in rocks means that the formation factor cannot be expressed simply as a function of porosity and
connectivity in a theoretically rigorous way.
Archie examined the way that the formation factor changes from rock to rock, and noticed that the
following rule commonly holds true
Equations (17.9) and (17.10) are often combined and called Archie’ s first law. The resulting equation
is
R0 = Rw ф-m (17.11)
The cementation index is the factor that describes the increase in resistivity that results from the
insulating mineral grains forcing the current to take tortuous pathways through the conducting fluid.
The cementation factor has a theoretical value of unity for uniform pores that penetrate the rock
directly from one side of the sample to the other (i.e., direct tubes of pore space), and is zero for a rock
with 100% porosity (i.e., no grains to get in the way of the fluid flow). No other values of the
cementation factor are able to be defined in a purely theoretical way in rocks due to the complexity of
the way that pore spaces are arranged.
In real rocks the cementation index usually varies between 1.0 and 3.0.
Values between 1.0 and 1.4 are associated with igneous and metamorphic rocks that contain fractures.
Fractures are a form of porosity that is localized and well connected, and hence approximates to the
situation where we had uniform tubes of porosity going through the sample.
Values between 1.4 and 2.0 are found in sandstones, with the higher values found in more
consolidated sandstones, where the current flow paths are more tortuous.
Values between 2.0 and 2.6 are typical for carbonates, and represent a greater degree of tortuosity in
the current flow that is found in carbonates because much of the porosity in carbonates is unconnected
(e.g., vugs).
Both the formation factor and the cementation exponent can be measured on core plugs in the
laboratory. This is done in the following way for a single plug.
Figure 17.8 shows a formation factor-porosity cross-plot for various values of the cementation
exponent.
It should be mentioned that there are a range of equations used in the oil industry to calculate the
formation factor. Archie’s equation is the most flexible, and the others are simply specific cases of it
that have been shown to work empirically for a given type of rock. In all cases the equations have been
derived from fitting a best line to a set of real data. The empirical relationships this provides, therefore
includes values of a which are non-unity. Also, one should be extremely careful to only apply the
equations when you are sure that they are valid. This often means going back to the original papers to
see what type of rocks the equations have been derived from. By comparison the use of Archie’s
equation with known cementation exponent is more reliable, and should ALWAYS be used if the
cementation exponent is known.
The Humble Formula. This is applied to soft formations and to clean sandstones with a sucrosic
texture.
(17.2)
(17.3)
The Low Porosity Carbonate Formula. Valid for low porosity clean carbonates with no fracturing.
(17.4)
Rt = I Ro (17.15)
The constant of proportionality I is called the resistivity index and describes the effect of partial
desaturation of the rock.
The resistivity index therefore varies between unity and infinity depending upon the degree of
saturation of the rock.
Archie observed that the following relationship exists empirically for sandstones
Again, the last two equations can be combined into a form which is usually referred to as Archie’s
second law
(17.17)
The saturation exponent normally has a range of values from 1.8 to 2.0, however much lower and
much higher values have been found.
The value of the saturation exponent can be obtained from laboratory experiments on core samples.
The procedure is as follows for a single core sample:
• Follow the procedure to measure the resistivity of a sample saturated completely with a conductive
fluid as outlined in the last section.
• Replace in a step-wise manner some of the conductive fluid in the rock with a non-conductive
fluid (e.g., gas) allowing for equilibrium to be attained at each step.
• Measure the resistivity of the sample when equilibrium is attained at the end of each step.
• Calculate from measurements of the evolved fluids and prior knowledge of the pore volume of the
sample, the conductive fluid saturation in the sample at the end of each step.
• Rearrange and apply Eq. (17.15) to calculate the saturation index.
• Rearrange and apply Eq. (17.17) to calculate the saturation exponent.
These experiments are extremely time consuming as one has to wait a long time for the samples to
come to equilibrium.
As before, a mean saturation exponent can be calculated from such measurements on a suite of cores.
In this case the resistivity index I is plotted against the water saturation Sw, again on log-log paper
(Fig. 17.9). The result is a straight line intersecting I=1 when Sw=1, and with a gradient equal to –n.
The two equations for each of the Archie laws can be combined into one controlling equation.
Combining Eq. (17.11) and (17.17) gives
(17.18)
We are ultimately interested in calculating the water saturation, Sw, so we rearrange the equation to
give
(17.19)
Table 17.1 summarizes the sources of the parameters that go into this equation to calculate the water
saturation.
Parameter Source
Rt • Deep investigation resistivity tool
Rw • From SP log
• Calculated from water zone
• Measured on FT water sample
ф • Acoustic tool
• Formation density tool
• Neutron tool
m • Measured in laboratory
• Guessed
n • Measured in laboratory
• Guessed
In attempting to reduce errors we ensure that the five parameters in Eq. (17.19) are measured using
independent methods. Laboratory determined m and n values are the best ones to take, however early
in a reservoirs life these are not available, and so guesses are used instead.
Table 17.2 shows the propagation of errors in Eq. (17.19) for the calculation of water saturation.
Note that 20% underestimations and overestimations of m and n lead to an underestimation and
overestimation of the water saturation by 0.11 and 0.12 respectively. These are huge errors when
progressed through to the STOOIP calculation. Clearly there is a case here for doing m and n
determinations on core.
Note also that the same degree of underestimation or overestimation in Rt and Rw leads to no further
error as these errors cancel out.
The worst case scenarios for over and underestimation with 20% errors on the input parameters are
0.32 and 0.22 respectively. These errors will most probably be larger than the recoverable oil
saturation in the reservoir. Clearly, the parameters that go to calculate the water saturation must be
derived very carefully indeed.
(17.20)
Because in any given reservoir we can take Rw, m and n as constant, and because we will apply the
equation for selected values of Sw, Eq. (17.20) becomes
(17.21)
where B is a constant. The Hingle graph paper is designed such that the y-axis represents Rt-1/m so that
Rt can be entered directly in the plot. This implies that a different form of graph paper is needed for
each value of m. The x-axis on the Hingle grid is porosity on a linear scale.
Application. The use of the Hingle plot is as follows. For any given reservoir zone carry out the
following steps:
• Construct the 100% water saturation (Sw=1) line. The first point on this line is automatically
available, as Rt is infinite when ф = 0, and this point plots in the bottom left hand corner of the
Hingle grid. The second point is calculated with knowledge of Rw for the reservoir. Equation
(17.11) is used to calculate Ro knowing Rw for the reservoir, for the value of m relevant to the
Hingle grid, and at any value of ф (the higher the better for accuracy). For example, in Fig. 17.10,
the m value is 2, and if Rw = 0.4 ohm.m, we can say that at the arbitrary porosity of =0.2, the value
of Ro = 10 ohm.m. The Ro, ф point can be plotted on the grid and joined with the first point by a
straight line. This is the water line, and represents how Ro varies with porosity when the rock is
fully saturated with water.
• Other lines for partial water saturations can now be constructed. Their first point is always in the
bottom left hand corner of the Hingle grid because Rt is always infinite when ф = 0 no matter what
the water saturation. The second point is calculated from Eq. (17.17) at a given arbitrary porosity
assuming or knowing the value of n and calculating Rt from the relevant Ro, which is available
from the water line. For a particular partial saturation line (Sw = 0.5, say) the Rt, ф point can be
plotted on the grid and joined with the first point by a straight line. This is the Sw = 0.5 line, and
represents how Rt varies with porosity when the rock is 50% saturated with water.
• A fan of partial saturation lines can be constructed in this way, say for every 10% increment in
water saturation. A large number of porosity and Rt pairs are now extracted from the logs and
plotted on the graph. It is immediately obvious how much water saturation is present on average,
and the water saturation for particular points (relating to a particular depth) can be estimated from
the graph by interpolation between the iso-saturation lines.
This type of plot is rarely used today with the use of computers to analyse data.
(17.22)
(17.23)
So a plot of log Ro against log ф gives a straight line. The value on the y-axis is equal to log Ro when
ф=1, and the slope of the line is –m.
(17.24)
(17.25)
which is the same straight line as described by Eq. (17.23), with the same gradient, but with a parallel
shift equal to log I.
Application. The Pickett Plot plots the formation resistivity Rt against the porosity on a log-log scale.
The data form straight lines with a gradient equal to –m. Hence, the cementation exponent can be
calculated. If one has data in the water zone of the reservoir, Eqs. (17.22) and (17.23) hold true, and
the value on the y-axis when the line intersects ф=1, gives log Rw from which Rw can be calculated.
The line is called the water line.
If one has data in the oil-bearing zone, and the value of Rw is known, the value on the y-axis when the
line intersects ф=1, gives log I + log Rw from which I can be calculated if Rw is known. If the saturation
exponent is then known, we can use the I value to calculate the water saturation.
Alternatively, we can establish the water line and construct iso-saturation lines with the same gradient
that are offset from the water line by values of log I that represent increments in water saturation.
Plotting the formation resistivity and porosity values from logs on this plot then allows the mean
saturation in the reservoir to be judged, and particular values of water saturation at a given depth can
be calculated can be approximated by interpolation between the iso-saturation lines.
This plot is very useful and easy to produce with a modern log analysis cross plotting program.
We apply this equation to the reservoir zone (i.e., the uninvaded zone deep in the formation) and can
calculate the water saturation in the reservoir, and hence calculate the amount of oil in the reservoir.
However, we can also apply the equation in the flushed zone if we are using water based muds. Before
invasion, we had a water saturation Sw and an oil saturation So. In the flushed zone after invasion, the
water has been replaced with mud filtrate, and some of the oil has also been replaced by mud filtrate.
The oil that has been replaced by mud filtrate is the mobile oil, i.e., that oil which was able to be
pushed out of the way by the invading mud filtrate. We want to calculate the saturation of oil which
the mud filtrate was able to push further into the formation, because if it was mobile enough to be
moved in that way, the likelihood is that it will be mobile enough to be produced easily from the
reservoir.
(17.26)
Hence,
(17.27)
Here Rxomf is the resistivity of the formation flushed zone containing nothing but 100% mud filtrate,
i.e., invasion has replaced all the oil and water, while Rxoor is the resistivity of the flushed zone
containing residual oil.
The value Sxo is the saturation of water based mud filtrate in the invaded zone. Hence the saturation of
oil that was moved by the invasion is (Sxo - Sw), and the volume of moveable oil per unit volume of
rock is (Sxo – Sw). We can use this modified value of moveable hydrocarbons in the reservoir to
calculate a lower, and more accurate value for moveable STOOIP by inserting (Sxo – Sw) in place of (1
– Sw) in Eqs. (1.2) and (1.4).
The spontaneous potential log (SP) measures the natural or spontaneous potential difference
(sometimes called self-potential) that exists between the borehole and the surface in the absence of any
artificially applied current. It is a very simple log that requires only an electrode in the borehole and a
reference electrode at the surface. These spontaneous potentials arise from the different access that
different formations provide for charge carriers in the borehole and formation fluids, which lead to a
spontaneous current flow, and hence to a spontaneous potential difference. The spontaneous potential
log is given the generic acronym SP.
• Correlation.
The log has a low vertical resolution, and is always recorded in the leftmost track of the log suite,
together with the GR log.
It is very important to recognize that this log has no absolute scale – only relative changes in the SP
log are important. This is reflected in the design of the log header, which shows only a bar that
represents a change of, say, 10 mV.
18.2 Principles
There are three requirements for the existence of an SP current:
• A conductive borehole fluid (i.e., a water based mud, there is no SP in oil based muds).
• A sandwich of a porous and permeable bed between low porosity and impermeable formations.
• A difference in salinity between the borehole fluid and the formation fluid, which are the mud
filtrate and the formation fluid in most cases. Note, however, that in some special cases an SP
current can be set-up when there is no difference in salinity, but where a difference in fluid
pressures occurs.
The origin of the spontaneous potential has four different components. These are shown in Fig. 18.1.
The spontaneous potential is composed of contributions that are electrochemical (arise from electrical
interactions between the various chemical constituents of the rocks and fluids), and electrokinetic
(arise from the movement of electrically charged ions in the fluid relative to the fixed rock).
1. The diffusion potential (sometimes called the liquid-junction potential). This potential exists
at the junction between the invaded and the non-invaded zone, and is the direct result of the
difference in salinity between the mud filtrate and the formation fluid.
Assume that the formation fluid is more saline than the mud filtrate for a moment, and that the
only dissolved ions in the system are Na+ and Cl- , as NaCl. The chloride ions have a higher
mobility than the sodium ions. When the two fluids come into contact across the interface
between the invaded and non-invaded zones, diffusion will occur. Ions from the high salinity
formation fluid will diffuse into the invaded zone to try to balance the salinities out. The chloride
ions are more mobile and so more of them diffuse into the invaded zone than sodium ions. The net
result is a flow of negative charge into the invaded zone, which sets up a charge imbalance
(potential difference) called the diffusion potential. The diffusion potential causes a current to flow
(from negative to positive) from the invaded zone into the non-invaded zone. This scenario is
illustrated in Fig. 18.2 for an analogue system, and is applied to the borehole environment in Fig.
18.3.
Of course, if the mud filtrate has a higher salinity than the formation fluid, the same argument
applies but in reverse, and leads to a reverse diffusion potential and current flow.
The same arguments also apply for more complex fluid compositions because some ions always
have a greater mobility than others.
For NaCl solutions at 25oC, the diffusion potential, Ed, is given by; Ed = -11.8 1× log(R1/R2), where
R1 is the resistivity of the diluter solution, and R2 is the resistivity of the more saline solution.
- +
Cl- Na+
Na+
Cl-
Na+
Cl-
Low NaCl High NaCl
Na+ Conc.
Conc. Cl-
Na+
Cl-
Na+
Rmf Cl- Rw
Shale
Cl-
Rmf Rw
Low Salinity High Salinity
Na+
Shale
2. The membrane potential (sometimes called the shale potential). This potential exists at the
junction between the non-invaded zone and the shale (or other impermeable rock) sandwiching
the permeable bed. These beds are usually shale, and the argument that follows applies mainly to
shales, but is also valid to a less extent for other low permeability rocks.
Shales have the property that they can preferentially retard the passage of anions. This is called
anionic permselectivity or electronegative permselectivity and is a property of membranes. It is
due to an electrical double layer that exists at the rock-fluid interface, and that has the ability to
exclude anions from the smaller pores in the rock (sometimes called anion exclusion). The
strength of this effect depends upon the shale mineralogy, the fluid concentration and the fluid
pH. Most other rocks exhibit the same behaviour but to a lower degree for geologically feasible
fluid concentrations and pHs, but cationic permselectivity is possible, if rare. Most subsurface
shales are such efficient anionic permselecting membranes that they repel almost all anions (say,
chloride ions). This results in the shale being more positive than the non-invaded zone, and hence
there is an electrical membrane potential, which causes current to flow from the invaded zone into
the shale (and hence borehole). This scenario is illustrated in Fig. 18.4 for an analogue system,
and is applied to the borehole environment in Fig. 18.5.
For NaCl solutions at 25oC, the membrane potential, Em, is given by; Ed = 59.1 5× log(R1/R2),
where R1 is the resistivity of the diluter solution, and R2 is the resistivity of the more saline
solution.
+ -
Na+ Cl-
Cl-
Na+
Cl-
Na+
Low NaCl High NaCl
Conc. Cl- Conc.
Na+
Na+ Cl-
Shale
Barrier
Rm Na+ Cl- Rw
Anions excluded
Shale
Rw
High Salinity
Shale
The total electrochemical component of the SP at 25oC for NaCl solutions is therefore;
(18.1)
for our situation where the mud filtrate is lower salinity than the formation fluid.
(18.2)
where; K is a coefficient that depends upon temperature, Aw is the activity of the formation water, and
Amf is the activity of the mud filtrate. This relationship breaks down for very saline solutions (Rmf <
0.08 Ωm), when corrected formation and mud filtrate resistivities must be obtained from charts. In this
circumstance Eq. (18.2) is rewritten as
(18.3)
where Rmf and Rmfe are corrected resistivities obtainable from charts.
The electrokinetic contribution, itself, consists of two effects, which are usually very small and act in
opposite ways such that they cancel each other out. These contributions depend upon fluid flow, and
hence are larger when there is a substantial difference in pressure between the borehole and the
formation. Thus, these contributions may be significant for depleted and under-pressured reservoirs
where the differential pressure is high (>500 psi). The contributions also depend upon the
development of an electrical double layer at mineral surfaces, which is larger for low salinity fluids.
Hence, these contributions are also more important for fresh formation waters or mud filtrates.
1. The mudcake potential. This potential is produced by the movement of charged ions through the
mudcake and invaded zone in a permeable formation. Its size depends upon the hydraulic pressure
drop, and since most of this is across the low permeability mudcake, the great majority of
electrokinetic potential is also generated across the mudcake, with an insignificant amount in the
invaded zone.
The surface of clay minerals carry a negative charge when in contact with fluids of pH and
salinities that are geologically feasible. The surface negative charge attracts cations from the
fluid, which become adsorbed to the surface and are fixed. The now apparent positive surface
charge is balanced out by an excess of negative ions in the fluid close to the surface. This “layer” of
excess negative fluid close to the surface ensures global charge balance and is mobile. If there is
no fluid flow the situation is electrically balanced. However, if there is fluid movement due to
passage of mud filtrate through the mudcake the mobile layer moves setting up a streaming
potential. Since the mudcake is anionically permselective (like shales), the potential that arises is
always negative, with current flowing from the borehole into the formation through the mudcake.
2. The shale wall potential. This potential is the same in origin to the mudcake potential, but
applies to the flow of fluids from the borehole into shale formations. It is usually very small
because the flow into impermeable shales is small. It also acts to set up a current flow into the
formation. We will see how this tends to cancel out the mudcake potential.
The total electrokinetic potential is E k = E mc + E sw, and because E mc and Esw have the same
polarity, the value of Ek is the difference between their absolute values, i.e., Ek = |Emc| + |Esw|.
• First assume that point A in the borehole has some unknown potential relative to the surface Eo.
• The mudcake potential Emc induces a current flowing into the formation through the mudcake.
Therefore at point B, the potential is Eo + Emc and current has flowed from A to B.
• The diffusion potential Ed across the interface between the invaded and non-invaded zones induces
a current flowing from the invaded zone into the non-invaded zone. Therefore at point C, the
potential is Eo + Emc + Ed and current has flowed from A through B to C.
Figure 18.6 Combination of the electromotive components of the spontaneous potential for the
formation water more saline than the mud filtrate.
• The membrane potential Em across the interface between the permeable formation and the shale
- - non-invaded zone. Therefore at point D,
above it induces a current flowing into the shale from the-
the potential is Eo + Emc + Ed + Em and current has flowed --from A through B and C to D.
-
-
-
-
Geoff Page 2010 Page 227
Petrophysics MSc Course Notes Spontaneous Potential Log
• The shale wall potential Esw induces a current flowing into the shale from the borehole. This
current counteracts the current flow set-up in the previous steps. Therefore at point E, the potential
is Eo + Emc + Ed + Em - Esw and the current has flowed from A through B and C to D, and has been
slightly reduced there by the small countercurrent due to the shale wall potential. The overall
effect is for a net current to flow from A through B, C and D to E.
• The overall effect is for point E in the borehole opposite the shale wall to have a more positive
potential than that at point A opposite the permeable formation. Hence there will also be a current
flow in the borehole between the borehole opposite the shale beds and the borehole opposite the
permeable bed to close the loop.
The arrows in Fig. 1 8.6a show the flow of currents and the black charge symbols indicate the local
relative potential due to each of the contributions to the SP. The grey charge symbols indicate the
overall potential set-up by the combined process. Figure 18.6b shows the summation of the potentials.
Note, since the value of Eo is arbitrary, there is no absolute value of spontaneous potential – what
matters is the relative change in spontaneous potential.
So, is the SP opposite a permeable formation always less than that opposite the shale above (or below)
it? NO. This is for the particular case where the formation fluid is more saline than the mud filtrate.
If the formation fluid is less saline than the mud filtrate, the opposite applies, and the SP opposite the
permeable formation is GREATER than that opposite the shale above (or below) it. This situation is
shown in Fig. 18.7, and described in the step-by-step list below.
• First assume that point A in the borehole has some unknown potential relative to the surface Eo.
• The mudcake potential Emc induces a current flowing into the formation through the mudcake.
Therefore at point B, the potential is Eo + Emc and current has flowed from A to B. This is the same
as in the previous case because nothing has changed locally. The mud filtrate is still passing
through the mudcake and sets up the same potential difference.
• The diffusion potential Ed across the interface between the invaded and non-invaded zones now
induces a current flowing from the non-invaded zone into the invaded zone (i.e., the opposite way
to the previous case). Therefore at point C, the potential is Eo + Emc - Ed and current has flowed
from C to B. Since this current is generally larger than that flowing from A to B in the opposite
direction due to the mudcake potential, the net current flow is from C through B to A.
• The membrane potential Em across the interface between the permeable formation and the shale
above it now induces a current flowing out of the shale and into the non-invaded zone (i.e., again
the opposite way to the previous case). Therefore at point D, the potential is Eo + Emc - Ed - Em and
the net current has flowed from D through C and B to A.
• The shale wall potential Esw induces a current flowing into the shale from the borehole as it did in
the previous scenario. This is because the local situation has not changed – mud filtrate is still
flowing into the shale and setting op the associated streaming potential. This current now does not
counteract the current flow, but contributes to it. Therefore at point E, the potential is Eo + Emc - Ed
- Em - Esw and current has flowed from E through D, C and B to A, and has been slightly reduced
by the small countercurrent due to the mud cake potential between B and A.
• The overall effect is for point E in the borehole opposite the shale wall to have a more negative
potential than that at point A opposite the permeable formation. Hence there will also be a current
flow in the borehole between the borehole opposite the permeable bed and the borehole opposite
the shale beds to close the loop.
The arrows in Fig. 18.7a show the flow of currents and the black charge symbols indicate the local
relative potential due to each of the contributions to the SP. The grey charge symbols indicate the
overall potential set-up by the combined process. Figure 18.7b shows the summation of the potentials.
Note, if the mud filtrate and the formation fluids have the same or similar salinities, there is no change
in the SP Log between shaly and permeable formations.
Figure 18.7 Combination of the electromotive components of the spontaneous potential for the
formation water less saline than the mud filtrate.
Only relative changes in potential are measured because the absolute value of the SP is meaningless.
Changes of the order of 50 mV are typical. For the log to be good, a good earth is necessary, which is
often a metal spike driven 1 m into the ground.
As stated above one of the requirements for measuring an SP is a connection to a good surface earthing
point. This is not a problem in wireline logs, where the logging cable provides the connection, however
this is a BIG problem for LWD logs where there is no cable. Currently the SP is not available from
LWD data.
The SP log is difficult to run offshore because (i) a good earth is difficult to find, and (ii) the amount of
electrical noise on board a rig often causes problems for accurately measuring signals that
commonly change by less than a millivolt. Sometimes the riser is used as the earth, but often the riser
is in connection with the rig, and therefore a source of electrical noise. Sometimes the rig legs or an
anchor chain are used, which is a big mistake because (i) the legs of oil platforms are given a specific
electrical potential to help combat the effects of corrosion, and (ii) the operation of sea waves on the
rig legs induces potentials in the rig legs that give rise to wavy patterns in the recorded SP log which
cannot be removed. In shallow waters the best method to record an offshore SP is to use an electrode on
a long cable that is thrown into the sea, as far from the rig as possible, and lies on the seabed. The
electrode consists of a simple small lump of lead, and is often referred to as an SP “fish”.
In reading the SP log it is best to first define a shale base line. This is the typical SP level for shales
and can be found by comparing the SP log with the GR log response. Permeable formations will then
have excursions of variable intensity to the left or right of this line, depending upon the relative
salinities of the formation water and the mud filtrate. It is useful to know the salinity or resistivity of
the mud filtrate from the log header, if available, as this will indicate whether the formation water is
likely to fall at a higher or lower salinity. For example, if the mud filtrate is known to be very fresh,
the likelihood is that that the formation water will be saltier, and the SP will likely kick left.
Bed resolution is bad, and one would not expect it to show beds less than about 20 times the borehole
diameter.
The recorded SP log represents the potential drop in the borehole, only. To use the SP curve
quantitatively, a value for the total potential drop around the circuit must be derived. This is called the
static spontaneous potential (SSP). This value may be derived from correction charts. However, a
direct reading of the SSP may be obtained directly from the SP log opposite thick, clean, shale-free,
100% water-bearing formations. The SSP is the value in millivolts of the difference between the SP
log at the shale base line and that in the centre of the thick clean formation, as shown in Fig. 8.11.
SSP Value
(mV)
The SP deflection obtained for homogeneous shaly formations or thin shaly beds after correction for
bed thickness is called the pseudo-static spontaneous potential (PSP). The SSP is the value in
millivolts of the difference between the SP log at the shale base line and that in the centre of the thick
homogeneous shaly formation, or a thinner bed if a bed thickness correction has been carried out.
If there is a proportion of shale in the permeable bed, the SP deflection is reduced from what it would
be if the bed were clean and contained the same fluids. Hydrocarbon saturation also decreases SP
deflections.
The present understanding of the electrical effect of clays and shales in reservoir rocks depends largely
upon the concept of cation exchange capacity, Q v , which is the cation concentration in
milliequivalents of exchange sites for sodium ions per cubic centimetre of pore volume. Laboratory
investigations have used this concept to develop a method for calculating the formation water
resistivity from SP log data in a way that takes account of the shaliness of a formation.
To use this method a value of Qv is needed from the shale beds above or below the formation of
interest, and a value of Qv for the formation of interest. These values are normally obtained from
standard chemical methods carried out in the laboratory upon cores or sidewall cores. More
information concerning Qv is given in Chapter 20.
The SP log can be used to follow facies changes. However, it has been largely replaced by the GR log,
which has a higher resolution and is more reliable.
18.7.4 Calculation of Rw
This is one of two quantitative use of the SP log. However, it is extremely useful when no formation
water samples or water-bearing sands are available to otherwise obtain Rw from during an analysis for
OOIP.
This is the quickest and most common method, and does not require knowledge of Qv. As the method
ignores the complicating effects of electrokinetic potentials and the possible presence of clay, it is to
be used with caution as errors in Rw will translate into errors in OOIP that may represent tens of
millions of dollars.
[1] Read off the difference in millivolts between the shale base line and the SP curve in the centre
of the formation of interest.
[2] Correct the SP reading from [2] for bed thickness using the correction chart shown in Fig.
18.14. Enter the bed thickness, read off the correction factor, and multiply the correction factor
by the reading from [2]. This gives the SSP for the depth concerned.
[3] Enter this value into the correction chart in Fig. 18.15 y-axis, and intersecting the relevant
temperature curve, read off the value of the ratio Rmfe/Rwe from the x-axis.
[5] Now go back to the initial data provided by the log header to find Rmf @ 75oF.
• If Rmf @ 75oF (24oC) > 0.1 Ωm, correct Rmf to the formation temperature using Fig. 18.13
as in [1], and use R mfe = 0.85× R mf
• If Rmf @ 75oF (24oC) < 0.1 Ωm, determine the value of Rmfe at the formation temperature
using the correction chart shown in Fig. 18.13 as in [1].
[6] With the known value of Rmfe at the formation temperature, calculate the value of Rwe at the
formation temperature from the Rmfe/Rwe ratio calculated in [4].
[7] Determine the value of Rw at the formation temperature from calculated Rwe at the formation
temperature using the correction chart shown in Fig. 18.16.
This is a complex method that is the most accurate we have. It accounts for both electrokinetic
potentials and the effect of the shaliness of the formation, but does not account for the possible
presence of hydrocarbons.
[1] Read off the difference in millivolts between the shale base line and the SP curve in the
centre of the formation of interest.
[2] Correct the SP reading from [1] for bed thickness using the correction chart shown in
Fig. 18.14. Enter the bed thickness, read off the correction factor, and multiply the
correction factor by the reading from [1]. This gives the SSP for the depth concerned.
[3] Determine the mudcake potential Emc for your particular mud (Fig. 18.17).
[4] Determine the shale wall potential Esw for the correct pressure differential between the
borehole and the formation using Esw = ∆P/100, where the pressure difference is in psi.
[5] Derive the total electrokinetic contribution to the SP using E k = SSP + |E mc | - | E sw | at the
appropriate formation temperature.
[7] Obtain the mud filtrate salinity in g/l NaCl equivalent from the mud filtrate resistivity value.
[8] Select the correct “Qv – shale” chart (Figs. 18.18 and 18.19) depending upon the value for your
formation in meq/cm3 . Enter the mud filtrate resistivity on the x-axis, project vertically
to intersect the appropriate Qv reservoir curve, project horizontally to read the “E” value.
[9] If Ek is negative, subtract the absolute value of Ek calculated in [6] from the E value from [8].
[10] If Ek is positive, add the absolute value of Ek calculated in [6] to the E value from [8].
[11] Enter the new value into the y-axis of the chart, projecting horizontally to the appropriate Qv
reservoir curve and down to the x-axis again. read off the formation water salinity.
[12] Convert the formation water salinity to formation water resistivity using a chart (Fig. 18.13).
(18.4)
PSP = SP log read in a thick homogeneous shaly sand zone, SSP = SP log read in the thick clean sand
zone.
This assumes a linear mixing relationship between the SP log and shale volume, and has no theoretical
basis. It probably overestimates the shale volume.
Figure 18.16 Correction chart for Rw from Rwe for various formation temperatures.
Figure 18.17 The electrokinetic mudcake potential for treated and untreated drilling muds as a
function of mud resistivity for various fluid pressure differentials (mud pressure minus formation fluid
pressure). This data has been derived from laboratory measurements.
Figure 18.18 Smit method correction chart for obtaining water resistivity from SSP for various
formation temperatures. This chart is for a shale Qv = 1 meq/cm3, and plots electrochemical
potential against equivalent NaCl concentration in g/l.
Figure 18.19 Smit method correction chart for obtaining water resistivity from SSP for various
formation temperatures. This chart is for a shale Qv = 4 meq/cm3, and plots electrochemical potential
against equivalent NaCl concentration in g/l.
Electrical logs are perhaps the most important tools available to a petrophysicist. This is because they
provide a method for calculating the water saturation, upon which calculations of STOOIP are based.
They were also some of he first logs to be used, with Marcel and Conrad Schlumberger testing out an
electrical log for the first time in 1927 in the Pechelbronn field, France.
These first measurements were continuous recordings using 2 or 3 electrodes and a direct current. It
was discovered that high quality recordings of apparent resistivity could be obtained under favourable
conditions of small diameter boreholes, high mud resistivities and shallow invasion in thick reservoirs.
These early tools are called electric logging tools.
The development of electrical tools has henceforward been intense. There are now tools that can cope
with extremely highly resistive muds (oil-based muds or gas as the borehole fluid), which rely upon
electromagnetic coupling and an induced alternating current (induction logs). The induction log
actually measures conductivity, and hence is sometimes called the conductivity log. The modern tool
for measuring resistivity in high salinity (low resistivity) muds is the laterolog, which focuses its
current into a thin sheet to improve vertical resolution and penetration depth. The laterologs measure
resistivity in the conventional sense, and are usually referred to as resistivity tools. Both the induction
logs and the laterologs come in different types, which are sensitive to different depths of penetration
into the borehole. Hence resistivity determinations for the invaded, partly invaded and undisturbed
rock zones can be measured. In addition, there is a range of smaller electrical devices (micro-
resistivity tools), which are designed to measure the resistivity of mudcake. There are also Array Logs,
which are state of the art tools, and electrical measurements are used at high resolutions (small scale) to
image the interior of the borehole electrically (image logs).
The electrical tools also have a number of qualitative uses, principle of which are (i) indications of
lithology, (ii) facies and electro-facies analysis, (iii) correlation, (iv) determination of overpressure,
(iv) determination of shale porosity, (v) indications of compaction, and the investigation of source
rocks.
Resistivity Logs
Deep
Shallow
0.1 105
Shale
Salt Porous
Water Sandstone
Shale
Fresh Porous
Water Sandstone
Shale
Tight
Sandstone
Shale
Gas
Oil
Porous
Salt Sandstone
Water
Shale
Shaly
Fining-Up
Sandstone
Clean
Shale
These logs will be discussed briefly because data from them may still be encountered when
reanalyzing mature fields. They also form the basis of “BKZ” logs still run today in Russia and
other Eastern Block countries.
Take an homogeneous and isotropic medium that extends to infinity in all directions. Now pass a
current from an electrode A in the medium to another B infinitely distant. We take the potential of
electrode A to be some value VA = V, and that of electrode B to be zero, VB = 0. The current will flow
radially (Fig. 19.2), and generate spherical equipotential surfaces with electrode A at their centre. A
third electrode M placed near A will lie on one of these equipotential surfaces, whose radius is r. If we
connect electrode M to electrode B through a potential measuring device (voltmeter), it will show the
value of the potential on the equipotential surface that passes through M, VM. The resistivity of the
material between A and M is the calculated as:
(19.1)
Here: ∆V = the potential difference between A and M, the factor 4πr is defined by the geometry of
the system (spherical symmetry), and it is assumed that A and M are close enough together for the
current I to be constant, even though it is spreading out with distance.
Current
Flow
Lines
M
r Equipotential
Spheres
Different types of resistivity tool have different geometrical factors that depend on their electrode
arrangements. These are calculated theoretically, and checked in tool calibration.
This theoretical scenario is the basis for the original tools: what are called the normal logging devices.
The distance AM = r is called the spacing. Two spacings were commonly used, a short normal
spacing equal to 16 inches, and a long normal spacing equal to 64 inches. The longer the spacing, the
greater the depth of penetration of the current into the formation, but the lower its vertical resolution.
Although the theory is developed for 3 electrodes, and this is how the first measurements were made, a
four electrode arrangement soon became standard. This allows the current flow circuit (the generator
circuit in Fig. 19.3) to be separated from the potential sensing circuit (the meter circuit in Fig. 19.3),
which provides better quality results. In this arrangement a constant known current is flowed from A
to B (or B to A), and the potential is measured between M and N. Electrode B and N are kept at a long
distance from electrodes A and M to provide quasi-infinite reference points for the current and
potential measurements.
B N
N
Impermeable Impermeable
Shale Shale
M
Permeable Permeable B
Formation M Formation
A
A
Impermeable Impermeable
Shale Shale
Fig. 19.3 The standard normal configuration. Fig. 19.4 The standard lateral configuration.
Another arrangement is possible, where electrodes A and B are placed close together with respect to
the distance between A and M. This is shown in Fig. 19.4, and was called the lateral configuration.
The LL7 has 7 electrodes. A constant current is emitted from the centre electrode. A bucking current is
emitted from the two far electrodes, and is automatically adjusted such that the two pairs of monitoring
electrodes are brought to the same potential difference. Then the current from the central electrode is
focused in a thin disk far out into the formation. The potential between one of the monitoring electrodes and
the potential at infinity is then measured, and knowing the current from the central electrode allows the
formation resistivity to be calculated providing the geometrical factor of the arrangement is known
(calculated theoretically and tested in the calibration of the device). This electrode arrangement produces
a thin disk of current that is confined between the two sets of measuring electrodes (32 inches apart).
The strongly focused beam is little affected by hole size, penetrates the invaded zone, and measures the
resistivity of the virgin formation, Rt. The vertical resolution of the LL7 is 3 ft. and the sensitivity is 0.2
to 20,000 Ωm.
Corrections for all of these tools for borehole size, effect of invasion and thin beds are available in the
form of charts from the logging companies. However, the first two of these corrections are usually
minor.
U
M2 M2 U
M
Fig. 19.6 The DLL electrode configuration in both the LLd and LLs modes.
In the LLd mode, the tool operates just like a LL7 tool but with the same bucking currents that are
emitted from the A1 electrodes also being emitted from the additional farthest electrodes, A2. The
result of this is to focus the current from the central electrode even more than was the case for the LL7.
In the LLs mode, the A1 electrodes emit a bucking current as they did in the LL7 device, but the A2
electrodes are set to sink this current (i.e., the bucking current comes out of A1 and into A2 electrodes).
This means that the bucking current must veer away from the pathway into the formation, and back
towards the tool A2 electrodes, and hence cannot constrain (focus) the current being emitted from the
central electrode as much. The overall result is that the central electrode current penetrates less far into
the formation before it dies away.
Both modes of the dual laterolog have a bed resolution of 2 feet, and a sensitivity of 0.2 to 20,000 Ωm.
To achieve this sensitivity both the current and voltage are varied during the measurement, keeping
their product (the power) constant.
The dual laterolog can be run either eccentered or centralized (best) to reduce the borehole effect on the
LLs. A micro-resistivity device, usually an MSFL or MLL, is generally run in combination with the
tool. Hence, this tool combination examines the resistivity of the formation at three depths of penetration
(deep, shallow, and very shallow).
The resistivity readings from this tool can and should be corrected for borehole effects and thin beds,
and invasion corrections can be applied using the three different depths of penetration. An example of
dual laterolog data is shown as Fig. 19.7.
Array Laterolog tools are now also available that work on similar principles, however they have a larger
number of depths of investigation, typically 4 or 5 measuring the invasion profile from Ri to Rt.
Examples of these are the Rt Explorer (RtEx – Baker Hughes), and the High Resolution Lateral Array
(HRLA – Schlumberger).
The spherically focussed log (SFL) has an electrode arrangement (Fig. 19.8) that ensures the current is
focussed quasi-spherically. It is useful as it is sensitive only to the resistivity of the invaded zone.
Formation
The tool is pad mounted, and the distance across the pads is also recorded, giving an additional caliper
measurement. The arms of the caliper are kept in the fully collapsed state when inserting the tool
into the hole. However, the log is recorded during this insertion to give a log of mud resistivity Rm with
depth and at the BHT. Figure 19.10 shows an example of the caliper showing beds clearly on both the
minilog and the caliper traces.
Fig. 19.10 An example of a ML log showing mudcake from both the caliper and ML logs.
19.6.2 The
Microlaterolog
The microlaterolog (MLL) is the
micro-scale version of the
laterolog, and hence incorporates a
current focussing system. The tool
is pad mounted, and has a central
button electrode that emits a
known measurement current M2
M1
surrounded coaxially by two ring-
shaped monitoring electrodes,
and a ring-shaped guard electrode
that produces a bucking current
Ao
as in the DLL (Fig. 19.11). The
A1
spacing between electrodes is about
1 inch. It is commonly run with the
Electrodes
DLL for the purpose of
measuring R XO.
Rubber Pad
Mudcake
The tool operates in the same way as the LL7. The focussed current beam that is produced from the
central electrode has a diameter of about 1 1/2 inches and penetrates directly into the formation. The
influence of mudcake is negligible for mudcakes less than 3/8 ” thick, and in these conditions RXO can
be measured. The depth of investigation of the MLL is about 4 inches.
The current beam emitted by this device is initially very narrow (1”), but rapidly diverges. It has a
depth of penetration of about 4” (similar to the MLL). The initial narrowness of the current beam
means that its sensitivity to mudcake is somewhere between the MLL and the PL, and is not
significantly affected by mudcake less than ¾ ” thick.
The sonde consists of 2 wire coils, a transmitter (Tx) and a receiver (Rx). High frequency alternating
current (20 kHz) of constant amplitude is applied to the transmitter coil. This gives rise to an
alternating magnetic field around the sonde that induces secondary currents in the formation. These
currents flow in coaxial loops around the sonde, and in turn create their own alternating magnetic
field, which induces currents in the receiver coil of the sonde (Fig. 19.14). The received signal is
measured, and its size is proportional to the conductivity of the formation. Clearly there will be direct
coupling of the transmitter coil and the receiver coil signals, however the directly coupled signals can
be differentiated by their 90 ºphase shift and excluded. Additional focussing coils serve to improve the
vertical resolution and depth of penetration of the tool.
Receiver
Secondary Amplifier
Current Causes
Signal in Rx Coils
Secondary Foucault
Current Induced in
Formation
Magnetic Field
Causes Secondary
Current Transmitter
Oscillator
The intensity of the secondary currents generated in the formation depends upon the location in the
formation relative to the transmitter and receiver coils. Hence there is a spatially varying geometrical
factor to take into account. Figure 19.15 shows two ground loops of secondary current induced by the
transmitter and sensed by the receiver. The actual signal recorded by the receiver will be the sum of all
ground loops in the space investigated by the tool. Figure 19.16 shows the sensitivity map for this
space for a homogeneous medium, showing that 50% of the total signal comes from close to the tool
(borehole and invaded zone) between the transmitter and the receiver.
Tx Coil
The 6FF40 induction-electrical survey log (IES-40, or IEL) is a 6 coil device with a nominal 40 inch
Tx-Rx distance, a 16 inch short normal device and an SP electrode. An example of its data is given
as Fig. 19.17.
These tools are rarely run today, however a lot of historical data of this type exists.
Fig. 19.22 Differences in bed resolution from different electrical tools (courtesy of Rider [1996]).
Fine bed structure requires very shallow reading tools in order to obtain sufficient bed resolution.
Changes in formation microstructure (texture) are best seen on logs that measure the invaded zone.
This is because the texture of the rock is a significant parameter controlling the replacement efficiency
of the formation fluids with the mud filtrate, and hence changes in the texture have a large effect on
the degree of invasion, that is picked up by these medium depth penetration logs.
If the conductivity from induction-type logs is displayed, the units are millimho per metre (mmho/m)
and the scale is usually 0 – 2000 mmho/m (note the SI equivalent of mmho/m is millisiemens per
metre, mS/m).
The ML log is usually plotted in Track 2 over a range of 0 – 10 Ωm for both the micro-normal and
micro-inverse curve. Separation in these indicates permeability (Invasion).
Array logs generally have 4-6 curves, presented in terms of resistivity over an appropriate log
scale. Figure 19.24 shows some typical log presentations for electrical logs.
Figure 19.25 shows the characteristics that are being looked for schematically (for a water based mud).
Fig. 19.25 The response of resistivity logs in formations with various fluids (recognition of
hydrocarbon zones).
• If all three curves are low resistivity, and overlie each other, the formation is an impermeable
shale, or, rarely, the formation is permeable and water-bearing but the mud filtrate has the same
resistivity as the formation water.
• If all three curves are higher resistivity than the surrounding shales, and overlie each other, the
formation is an impermeable cleaner formation (sandstone, limestone).
• If the shallow curve has low resistivity, but the medium and deep penetrating tools have a higher
resistivity that is the same (they overlie each other), the formation is permeable and contains only
formation water, or there is very shallow or very deep invasion.
• If the shallow curve has low resistivity, the medium as a higher resistivity, and the deep one has an
even higher resistivity (i.e., there is separation of the medium and deep tool responses), the
formation is permeable and contains hydrocarbons or fresh water.
Note 1: If the mud filtrate resistivity is constant, the effect is greater for formations with fresh
formation waters than those for saline formation waters (Fig. 19.26), and in the case of extremely
saline formation waters the deep resistivity in the formation can be smaller than that of the adjacent
shale beds.
Porosity
Lithology Porosity
Porosity
=0% =20%
Separation
Shale
Porous
Sandstone
Shale
Fig. 19.26 The behaviour of the resistivity log responses for different formation water salinities.
Note that the effect is greater for oil-based drilling muds than fresh water-based muds and saline
water-based muds in this order.
Some of the more advanced array type tools can now calculate the invasion profile of resistivity. An
example of this is shown in Fig. 19.19, where the resistivity of the formation is shown as a function of
depth into the formation in Track 3 as a colour coded map, and the interpreted flushed and transition
zones are given in Track 1 for the permeable intervals.
Figures 19.25 and 19.26 are given as examples only. This is because there are too many variable
parameters (mud filtrate resistivity, mudcake resistivity, formation water resistivity, water and
hydrocarbon saturations, lithology and porosity etc.) to be comprehensive. In general, the best
approach is to (i) see what type of tools have been used to create the logs, (ii) discover their
penetration depths, (iii) discover the type of drilling mud and the resistivity of the mud filtrate used in
the drilling, (iv) take account of information from other logs (caliper for mudcake, SP for permeable
zones, gamma ray for shale volume, and sonic, neutron and density tools for porosity), and then
interpret the resistivity curves from first principles!
information about the reservoir thickness, its area and porosity, and fluid compression factors in
STOOIP calculations to calculate the amount of oil in the reservoir (Chapter 1).
19.12.4 Correlation
Electrical logs are often used for correlation. The deep logs (6FF40, ILd, LLd etc.) are the best to use,
as they are sensitive to gross changes in the formations at a scale that is likely to be continuous with
other wells. However, it must be noted that the resistivity of a formation also depends upon formation
fluid pressure and formation water resistivity, which are non-stratigraphic variables, and changes in
them from well to well may confuse correlations.
Sands shales and carbonates have no characteristic resistivity as their resistivities depend upon many
factors including porosity, compaction, fluid resistivity, texture etc. However sequences of these rocks
can usually be traced by using invariant characteristics from bed to bed, such as a similar resistivity
reading as another set of beds or a characteristic roughness/smoothness of the curve within the bed.
Thus if the basic lithologies can be defined from other logs, the electrical logs help progress them
through the logged section. As the electrical logs are very sensitive to texture, they are extremely good
at discriminating between lithologies of different types providing the types can be defined by some
other log.
An example of this may include beds of shale separated by thin layers of siderite stringers and
concretions. The shallow penetrating electrical tools will pick out each thin siderite bed as a sharp
peak in resistivity.
However, electrical logs do provide characteristic responses for some lithologies (Fig. 19.28). The
most common are:
The beginning of overpressure zones can be seen by a sudden unexplained jump of the resistivity to
lower values in a uniform lithology. This is best observed in shales and is associated with the higher
porosity induced by the overpressure. Again, the best tools to use are the deep looking tools.
Source rocks may be recognized, and their maturity indicated by electrical logs. Immature sources
have little effect upon electrical logs, but this effect grows with the degree of maturity. Hence, it
is expected that it may be possible to calculate the TOC% from electrical logs under the
correct conditions.
Resistivity Logs
0.1 Deep 105
Shale Variable
Tight Limestone
80-6000
/Dolomite
Porous Limestone
/Dolomite
Variable
Coal 10-1,000,000
Variable
Halite
1 0,000-inf.
Anhydrite 1 0,000-inf.
Gypsum 1000
Variable
Chamosite-
<10
Rich Shale
Variable
Sandstone
with Pyrite <1
Variable
Shale Variable
Fig. 19.28 Characteristic resistivities from various lithologies recorded by resistivity logs.
All of the deep reading resistivity tools may be affected by formations that are dipping relative to the
wellbore. The deep measurements may be seeing some proportion of their signal from bedding which
intersects the wellbore above or below the tool measure point, which introduces a form of
filtering/smoothing which increases with increasing dip. This means that high resistivity reservoirs may
be averaged with low resistivity shales, reducing the apparent reservoir resistivity and increasing
apparent water saturations. Figure 19.29 shows a thick reservoir unit, with a true resistivity of 50 ohmm,
surrounded by shale of 1 ohmm (shaded area). The response of a 6FF40 induction log is shown for a
wellbore intersecting this reservoir at a relative dip of zero and fifty degrees. It can be seen that in the
deviated case the apparent measured resistivity drops to around 15 ohmm in this 50 ohmm bed, due to
the effect of the deviation.
These affects can generally be corrected by modern computer algorithms. They are minimal below
around 20 degrees dip, and above 70 degrees very difficult to correct.
A new generation of resistivity tools is now available which measures directional or Tensor resistivities.
Examples of these are the 3D Explorer (3DEX) from Baker Hughes, and the 3D AIT from
Schlumberger.
• Thin bedded formations, below the resolution of standard tools exhibit a horizontal to vertical
resistivity anisotropy (Figure 19.30).
• Fractures can introduce electrical anisotropy parallel and perpendicular to the fracture plane.
These tools and their interpretation are beyond the scope of this course.
The base configuration consists of a single transmitter, and two receivers (Figure 19.31).
The tools measure the phase change, and amplitude change or attenuation, of a high frequency
electromagnetic wave between two receivers. Both phase and attenuation are related to the resistivity of
the surrounding environment. These are calibrated to formation resistivity and may both be presented on
a standard logarithmic resistivity grid.
Most current tools have multiple spacings and operate at several frequencies. In general:-
The multiple, complex, responses of the many curves produced by these tools can be confusing.
Some contractors can now combine them all into a set of fixed depth of investigation/fixed vertical
resolution curves similar to the output of a wireline resistivity tool. An example of this is INTEQ’s
MPRTEQ processing.
XX400
XX450
Fig. 19.32 MPRTEQ processing. The centre track shows the typical rigsite presentation of LWD
phase and Attenuation measurements. The Right hand track shows the processed resolution
matched fixed depth on investigation data.
Geoff Page 2010 Page 270
Petrophysics MSc Course Notes Clay/Shale Effects on Porosity and Resistivity Logs
20.1 Introduction
The presence of clay minerals or shale in porous formations presents problems from the interpretation
of wireline logs. For most logs these problems have been discussed in the relevant chapter. The
problem is, however, especially bad in the interpretation of resistivity data, and also affects the
porosity logs. This is not only because the presence of clays and shale have a gross effect upon
resistivity values, but because such data effects the final calculated STOOIP for a given formation.
Even small amounts of clay can have a large effect, which is important because most reservoir sands
contain some degree of shaliness.
Note: The terms clay and shale are often used interchangeably by petrophysicists – but wrongly! It is
very common to see layers of shale in reservoirs, however a pure block of clay is very rare. Clays are
minerals, which are just one of the constituent components of shales, together with many other minerals.
The distribution of shale within porous reservoir formations can be classified into three groups:
The laminated and structural shales are part of the rock structure and are considered to have the same
porosity (water content) as adjacent shale beds because they were subjected to the same overburden
pressures. By contrast, the dispersed shales were only subjected to the formation fluid pressure, and
therefore have a higher porosity and water content.
• Total porosity.
• Effective porosity.
a) b)
Fig. 20.2 Porosity models. (a) Total porosity and Effective porosity. (b) Clay and shale definitions.
Figure 20.3 illustrates this method graphically. The graph is a triangle, whose apices are at the
following points:
Fig. 20.3 Porosity- Vsh correction chart for the density – neutron log combination.
The porosity obtained from routine core analysis is also the total porosity in the sense that the
procedures used to prepare the cores ensure that all oil is extracted from the cores, and the core is then
dried. Under these conditions all absorbed and interstitial water is removed from the shale. As a result,
there is often good agreement between routine core porosities and density log derived porosity data. If
there is a discrepancy, the core porosity is usually lower. This is because the methods used to measure
the porosity of the cores uses imbibed fluids (water, mercury or gas), and these fluids can only access
open pore spaces: there will always be some pore spaces that are isolated and cannot be accessed.
The difference between фt and фe is the fraction of water held by absorbing capillary forces in the shale
fraction of the rock. This amount of water can be appreciable depending upon the type and distribution
of the shale, and the salinity and pH of the formation fluid. The water is bound, and does not take part
in fluid flow. It is common, therefore, to ignore this water fraction and to prefer to use фe as a
representative porosity for the rock. However, one must be cautious in using the effective porosity in
calculations involving resistivity tools.
The NMR tool is sensitive to the difference between bound and free water (i.e., water associated with
(фt - фe) and фe, respectively), and can be used to derive these values directly, and then use them to
predict the permeability of the rock.
Fig. 20.4 Effect of shale/clay mode on the effective porosity of a shaly sand.
• If it is as a dispersed shale, the 10% additional shale take the place of porosity, so the new effective
porosity becomes 30 - 10 = 20%, and the point now plots at point D in the figure.
• If it is a laminated shale, the 10% additional shale reduce the overall effective porosity by 10%, so
the new effective porosity becomes 30 – (30× 10/100) = 27%, and the point now plots at point L in
the figure.
• If it is a structural shale, the 10% additional shale replaces volume for volume 10% of the original
quartz grains, and the effective porosity is unchanged at 30%. Hence the new point plots at point S
in the figure.
We can use these relationships to look at structural shale distribution patterns on the density-neutron
cross-plot.
Clay distribution patterns, indicating dispersed, laminated and structural shales can be analyzed with
the density and neutron log data (Fig. 20.5). In this figure:
Fig. 20.5 Density-neutron crossplot showing characteristic points for shaly sands and shales.
• “Shale pt” indicates wet shale – NOT Clay.
• Laminar shales fall on line I.
• Dispersed shales fall Line II, and when all pores are full Line III.
• Structural shales fall on line IV.
The resistivity of the reservoir rock is the sum of the conductivities of the clean sand layers and the
shale laminae:
(20.3)
or in terms of resistivity:
(20.4)
shale
shale
(20.5)
(20.6)
In practice the water saturation is found using an iterative method. The shale volume (Vsh) is derived
from the GR or SP logs.
There are many shaly-sand water saturation equations that have been proposed in the literature for
dispersed shale systems. The most common is the Simandoux equation, however it was not determined
from reservoir rocks but artificial clay slurries and should be used with great care:
(20.7)
If a shaly sand contains clays that have a low conductivity, or relatively low percentage of shale then
Archies equation is frequently the best to use. Waxman and Smits equation is the recommended
option for most other shaly reservoirs. It is more complex, but can easily be solved today with
computer interpretation programs.
Waxman and Smit’s equation relates the electrical resistivity of a water saturated shaly sand to the
water conductivity and the cation exchange capacity per unit pore volume of the rock, Qv (meq/ml). It
is this last factor that partly controls the size of the surface conduction. The method is independent of
shale distribution.
(20.8)
where: F* = the formation factor in shaly sand (determined at high fluid salinities).
Co = 1/Ro
Cw = 1/Rw
Cs = The specific surface conductance of the clays in the shale fraction.
The value of Cs = BxQv, where B is the equivalent conductance of the counterions as a function of
solution conductivity, Cw. The units of B are Ωcm/m. meq. Equation (20.8) is represented graphically
in Fig. 20.7.
Fig. 20.7 The conductivity of a shaly sand as a function of formation water conductivity.
(20.9)
This equation was developed and tested in the laboratory using the total porosity, and hence it should
be restricted to use with total porosity.
The concentration of clay counterions as defined by Qv can be determined in the laboratory on material
from core by a titration/conductivity method. This is done on crushed samples, so there is some
concern that the value may not be relevant for whole rocks. Clearly there will be more exchangeable
counterions possible in crushed samples, which have a higher surface area for reactions. Some whole
rock measurements exist, but are uncommonly used in the oil industry.
It has been noted that there exists in some areas a relationship between Qv and фt. If this is noted in a
reservoir (Fig. 20.8), the Qv value may be determined from the total porosity from log data. In shaly
sands Qv usually ranges between 0.01 and 2 meq/ml.
Empirical relationships between B and Rw have been established at different temperatures. Between
50oC and 200oC (120 – 390oF) the product BRw is not dependent upon temperature, and can be
represented as a function of water salinity alone, as illustrated in Fig. 20.9.
Fig. 20.8 Typical Qv -фt for reservoir intervals. Here Qv = 0.0006× фt -1.8.
In hydrocarbon-bearing formations the exchange ions associated with the clay become more
concentrated in the remaining pore water. This concentration Qv’ is related to Qv and the water
saturation Sw by Qv’= Qv/ Sw.
(20.10)
(20.11)
The equation is solved for Sw using an iteration process, computationally. A quick manual solution can
be obtained from the graph shown in Fig. 20.10, which is valid for the saturation exponents, n*=1.8
and 2.0.
(20.12)