Number 19, 1997 Middle East Well Evaluation Review
Number 19, 1997 Middle East Well Evaluation Review
36 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
Diagenesis
Physics revision position of all signals from individual pores,
each with a specific decay time constant.
Carbonates are much more susceptible to
For a particular decay time constant the
post-depositional alteration (diagenesis) NMR measurements are based on the signal magnitude reflects the part of the
than are siliciclastic rocks. Diagenesis can relationship between the ‘transverse total porosity having a certain ratio of
cause dramatic changes in carbonate pore relaxation time’ of protons, diffusing within volume to surface area, or ‘pore body size’.
size and shape. a pore, and the pore’s geometry. The
Typically each T 2 component of the
Compaction, carbonate aggregation of effects of transverse relaxation time are
received signal is presented in a spectrum of
recrystallized material and the formation of measured as the exponential decay time
signal amplitude versus T2 time constant
stylolites can all reduce porosity. Other constant (T2) of the magnetic signal pro-
(on a logarithmic scale) and called the ‘T2
diagenetic processes, however, such as dis- duced by the protons in the pore.
solution or reduction in crystal size (micriti- distribution’.
If there are no magnetic field gradients
zation), fracturing or dolomitization, may in the path of the diffusing protons, and The outputs from CMR well logs are
increase porosity. the natural tendency for proton relaxation derived from the ‘T2 distribution’. The area
While diagenetic processes have the within the bulk fluid is ignored, T2 is con- under the distribution gives the lithology-
potential to increase or decrease total trolled by surface relaxation. In a par- independent NMR porosity. Typically this
porosity, the general trend is for carbonate ticular pore T2 can be described by the will be spilt into free f luid volume and
porosity to decrease from high values (up relationship: capillary-bound fluid volume by applying a
to 60%) at deposition to much lower levels 1/T2 = r2S/V suitable T2 cutoff. In the latest development,
(typically less than 10%) after lithification where: T 2 is the exponential decay the CMR-200 tool, T2 distributions also con-
and burial to reservoir depths. Whatever its time constant of the signal produced by tain the very fast decay signals from clay-
effect on total porosity, diagenesis adds the protons, r2 is the ‘surface relaxivity’ of bound water, which can be differentiated by
further complexity to the primary pore the grain, S is the surface area of the pore applying a second cutoff at 3 msec (Figure
space of the carbonate rock. body and V is the pore body volume. 3.3). The logarithmic mean of the relaxation
The signal magnitude is proportional to time divides the area below the T 2 dis-
Textural classification the number of protons within the pore, tribution curve into equal parts, and
and the decay rate is related to the quantifies the mean pore body size.
All carbonate sediments are composed of The pore property that NMR tech-
number which contact the grain surface.
three textural elements: grains, matrix and niques measure is V/S. This ratio (and
When the pore fluid wets the grains,
cement. The grains are large skeletal therefore T2) scales with linear dimensions
protons throughout the pore undergo
remains or particles of chemical origin.
relaxation in this way. This does not occur for pores of the same shape, but varying
Matrix is very f ine-grained material, of
if the pore fluid is nonwetting. sizes. One linear dimension which strongly
chemical or biological origins, which is
Signals from the pores are additive. inf luences many impor tant reservoir
deposited as lime mud, and becomes fine-
The received signal is therefore the super- properties is the pore throat diameter.
grained calcite during lithif ication. The
cement is a crystalline material binding Figure 3.3 Bound fluid
matrix and grains together. The relative Definitions logging is just one of the
proportions of these components can be Signal Sands integrated interpretation
used to classify carbonate sediments. distribution solutions available with
the CMR tool. The tool
Dunham’s classification (Figure 3.2) is
can be combined with
based on the relative content of mud and the MDT* Modular
grains, with additional categories for Formation Dynamics
crystalline carbonates and ‘boundstones.’ Tester tool to verify any
This long-established classification runs indications of producible
hydrocarbons
from carbonate mudstones (lithified mud
matrix with less than 10% grains) to grain- 0.3 3.0 33 3000
stones that contain no mud.
T2(msec)
Within these carbonates the porosity
may take many forms depending on the
inherent fabric of the rock, and on the
types of processes that can occur during Clay Capillary Producible fluids
bound bound
and after deposition. On the smallest scale water water
fabric-selective porosity may exist within, or Total CMR porosity
between, individual particles.
CMR porosity>3msec
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 37
Carbonate sediments may have ary processes which may affect them, it is Fluids and carbonates
additional porosity on a larger scale due to not surprising that the convoluted pore
Initial oil migration from an underlying
processes such as shrinkage or animal space of a carbonate may be quite different
source rock into a reservoir occurs not by
burrowing after deposition. On the largest from that found in siliciclastic sediments.
gravity segregation, but by the action of
scale, porosity due to fracturing or dis- Carbonate porosity can be classif ied
capillary forces (Figure 3.5). Water from
solution of carbonate rocks can produce according to rules defined by Choquette
the potential reservoir is imbibed, by
‘pores’ up to the size of caverns. and Pray (Figure 3.4). This simple
capillary forces, into the source rock where
Given the wide range of origins for classification helps to bring some order to
it displaces oil droplets. The displacement
carbonate rocks, and the variety of second- the huge range of carbonate porosity types.
process is one in which the oil and water
move simultaneously in opposite directions
Fabric-selective Not fabric-selective Fabric-selective or not between the small pores of the oil-wet
source rock and the larger, water-wet pores
Interparticle Fracture Breccia of the reservoir.
Once the oil droplets have reached the
reservoir pores (but before they form a
Intraparticle continuous phase) they migrate upwards. As
these buoyant oil droplets rise they are
Channel Boring resisted by capillary pressures at the pore
Intercrystal throats and pass through only the larger
pore spaces and throats (Figure 3.6). If the
capillary pressure of the oil exceeds the
Moldic reservoir displacement pressure, oil will
Vug Burrow migrate through the largest pore spaces to
the top of the reservoir. The displacement
pressure is a function of rock wettability,
Fenestral
interfacial tension between the fluids and
pore throat size.
As the oil droplets accumulate at the top
Shelter Cavern Shrinkage
of the reservoir they coalesce to form a
continuous phase. The oil–water contact
Man-sized or larger pores (OWC) moves deeper as accumulation con-
Growth-
framework
with channel or vug shapes tinues and oil pushes the continuous water
phase out of the pores. The differential
pressure between the oil and water phases
Figure 3.4 Choquette and Pray’s classification of carbonate porosity types covers the full range of during this process is equal to the capillary
sedimentary, biogenic and diagenetic fabrics found in nonclastic rocks. (P Choquette and L Pray pressure forcing water from the pores
(1970). Geologic nomenclature and classification of porosity in sedimentary carbonates. AAPG
Bulletin (54) pp 207–250)
(Figure 3.7).
The differential pressure at a particular
pore throat is a function of oil column
Water-wet system height, oil specific gravity and formation
water specif ic gravity (controlled by
salinity). For example, if the specific gravities
Nonwetting phase of the oil and the water differ by 0.3, the
Water q Oil
capillary pressure 60 ft above the free
water level (FWL) will be 9 psi.
Rock Oil displacement of water from the
smallest pore throats requires the highest
Oil-wet system capillary pressures. As a result, the pressure
difference will push water out of the large
pores first and the smallest pores will resist
q water displacement. In addition, larger
Water Oil Wetting phase
pores will remain unswept near the FWL,
resulting in the characteristic water
Rock saturation profile (increasing with depth)
present in the oil–water transition zone.
After migration, some water may remain
Figure 3.5 Capillary forces draw water from the potential reservoir into the source rocks.
These curved interfaces between oil and water are found in capillary openings in water- above the level of the continuous water
N u m b e r 2 0 , 19 9 7
wet and oil-wet systems. The angle of contact between the fluids (q) varies as shown phase. This water is typically found at grain
contacts or in bypassed pockets of water-
wet rock where pores remain unswept.
38 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
These displacement processes occur in
siliciclastic and carbonate rocks. In carbon- Water
ates, however, the extreme pore-size rp
rt r
variation leads to initial reservoir
saturations which are usually far more com-
plicated than those in sandstone, or shaly Grain
sandstone, sequences.
r rt
Oil rp
The wonders of wettability
r
Wettability is the term used to describe the
ease with which a fluid can be separated
from a surface. It is a function of fluid and
2g 2g 2g 2g 2g 2g 2g
surface material properties. Pc =
r rt rp r
=
r rp rt
Most siliciclastic sediments are
deposited in lakes or seas and water is, Figure 3.6 Model of an oil globule (radius r) migrating through a pore throat
therefore, the initial fluid wetting the grains (radius rt) in a water-wet clastic rock. The g factor is interfacial tension, rp is the
of lacustrine and marine reservoirs. Even radius of the oil globule as it enters the pore throat, and Pc is capillary pressure
wind blown (aeolian) deposits are likely to
be flooded by water after deposition, and
Gradient
few siliciclastic reservoirs (with the notable Dp/Dh = roil g
exception of tar sands) have oil-wet grains.
Clearly this means that most reservoir Oil
rocks have had water in contact with their OilÐwater
contact
grains before oil migration. Regardless of
the preferred wettability of the rock it can
Transition Free water Capillary
only become oil-wet during or af ter
level pressure
migration, but only if oil comes into contact
with the grains. In carbonates, this contact Water
can occur through a diagenetic process,
Gradient
further complicating the picture. Dp/Dh = rwater g
Electrostatic effects on carbonate grain
surfaces may make these sediments hydro-
phobic. Experimental evidence indicates
that, in the presence of oil and water,
carbonate particles adsorb an organic coat-
Continuous Continuous
ing and over several million years it is poss- water phase oil phase
ible for carbonates to change from being
water-wet to being oil-wet. This change in
wettability will only occur in the large, oil-
bearing pores: smaller, water-filled pores will
remain water-wet.
The resulting oil coating around some Figure 3.7 Pressure in the continuous water phase above the free water level (FWL)
grains protects the rock from dissolution by exceeds the pressure in the oil by the capillary pressure. This pressure difference allows oil
acidic formation water, and some carbonate to enter smaller pore throats at greater heights above the FWL
sequences may have their porosity pre-
served by the presence of hydrocarbons. water saturation values. The micropores
also provide a continuous path for logging
current even when the water saturation in
Pores of all sizes
the pores is low. Under production, oil will
The fluid content of carbonates is closely flow from the macropores while the micro-
related to pore distribution. It is convenient, pores will retain their capillary-bound water.
therefore, to classify carbonate pore spaces Mesopores, in the size range between
by the behaviour of the fluids which they micropores and macropores, contain oil or
contain. water depending on oil column thickness
Macropores are easily filled with oil and, therefore, the capillary pressure acting
above the free water level, while micro- to flush water from the pores.
pores have retained water af ter oil
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 39
Saturation situations Carbonate interpretation has been Accurate predictions of wettability, per-
improved in this respect by using variable ‘m’ meability effects and ‘m’ and ‘n’ values
The wide variety of pore sizes and types values derived from additional logging across a reservoir are invaluable when con-
found in carbonate rocks, and the coexist- measurements such as those provided by structing reservoir simulations. It is not sur-
ence of oil-wet and water-wet pores, the EPT* Electromagnetic Propagation Tool. prising that studies of the geological
contradict the assumptions made in con- The nature of the porosity above the properties influencing wettability have con-
ventional petrophysical interpretations. The FWL controls wettability and saturation in cluded that the pore system within the
Archie equation, for example, with its con- this interval. Microporosity above the FWL rock, described by the pore throat size, has
stant cementation and saturation will be water-wet while the macroporosity a large inf luence on this impor tant
exponents, is likely to run into difficulties in may be oil-wet. In the presence of parameter.
carbonate sequences. mesopores, wettability will be variable and
The cementation exponent ‘m’ in the depend on oil column thickness. Reservoir simulation
Archie equation can vary with porosity and The saturation exponent ‘n’ in these
facies because the conductive paths in low- mixed porosity systems is a function of As discussed above, the volume of oil in
and high-porosity rocks are caused by dif- wettability and the saturation found in each place is intimately linked to the distribution
ferent types of pores. The challenge for the porosity type (Figure 3.8). It is very likely, of saturation throughout a rock. This in
interpreter is made even more difficult by therefore, that variations in ‘n’ will occur turn is affected by the wettability of the
that fact several pore types may coexist, within a carbonate sequence, at least across rock to oil and water during migration at
providing parallel paths for current flow. the oil–water transition zone (Figure 3.9). the beginning of the life of the reservoir.
The flow of hydrocarbons through a
reservoir during production is also
100
Oil governed by relative permeability to oil and
water, properties which are also controlled
Sw = 15% by wettability. Capillary-bound water
saturation is another factor af fecting
reservoir recovery.
Resistivity index
0 0 0 0
0.1 0.3 1 4 10 0.1 0.3 1 4 10 0.1 0.3 1 4 10 0.1 0.3 1 4 10
Pore throat radius (mm) Pore throat radius (mm) Pore throat radius (mm) Pore throat radius (mm)
N u m b e r 2 0 , 19 9 7
40 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
NMR to the rescue tional logs are unable to predict zone pro- indistinguishable on a CMR log from large
ductivity, and only NMR techniques can intergranular pores. It may be necessary to
Carbonate reservoirs provide a challenge identify the best intervals for perforation. use additional logs to quantify the noncon-
for geologists, petrophysicists and reservoir nected pore space in these cases.
engineers alike. Geoscientists continually Two relationships can be used to con-
encounter complex relationships between
Facies mapping
vert from CMR relaxation time to per-
permeability, irreducible water volume, Carbonate facies can be identified and meability. These are the Schlumberger-Doll
residual oil saturation, wettability and classified as much by porosity size and type Research (SDR) relationship
capillarity. Wettability can be related to as by grain composition. The CMR tool pro-
rock type and pore system, permeability to vides accurate information about carbonate kSDR = A(fCMR)4 (T2log mean)2
irreducible water, and so on. Within all of pore systems and this can help to identify
these relationships, however, there is a carbonate facies from logs. Cutoffs which and the Timur (BWP) method which is
common dependence on the pore throat differentiate pore sizes within the total based on irreducible water volume.
size in the reservoir rock. porosity also differentiate carbonate facies.
The CMR measurement can help all three For example, high permeability grainstones kBWP = A(fDN)4 [(fDN – BVirr) / BVirr]2
disciplines and, when used in combination can be distinguished from low permeability
with existing measurements, provides a fresh wackestones or mudstones by the smaller T2 Where k BWP is bulk water pass
insight into carbonate reservoirs. of the latter indicating smaller total permeability, fCMR is CMR porosity, fDN is
porosity:pore size ratio. density–neutron porosity and BVirr is bulk
volume irreducible water.
Porosity
Each method requires a different cali-
Carbonate interpretation benefits from the
Permeability
bration constant ‘A’, which can be deter-
most basic of NMR outputs: lithology- Reservoir permeability is governed by the mined graphically from experimental data
independent ef fective porosity. This pressure drop across pore throats as fluids (Figure 3.10), and formation porosity.
measurement complements the density and flow through them. Permeability is there- In the SDR relationship carbonates will
neutron logs where matrix density is fore controlled by the size of the usually have a value for A that is less than
uncertain (for example dolomitic lime- reservoir’s pore throats – a variable that 0.1 while in sandstones the value for A is
stones) or where neutron porosity is over- can be inferred from the NMR measure- generally around 4. This difference is due to
whelmed by environmental effects. ment of pore body size. This continuous the lower surface relaxivity of carbonates
NMR total porosity in simple carbonates permeability log is one of the most import- compared to sandstones. It may also reflect
compares favourably with other porosity ant benefits of reliable NMR measurements. differences in wetting fluid affecting the
tools, while the ability of NMR methods to NMR-derived permeability relies on the relaxation at the grain surface.
detect free fluids and to measure free fluid existence of a well-defined relationship In the Lower–Middle Cretaceous lime-
porosity can add greatly to the between pore body size and the diameter stones of nor thern Oman, where the
geoscientist’s understanding of complex of pore throats. Unfortunately, this relation- lithology is predominantly packstone to
carbonates. Subdivision of NMR porosity ship is less predictable in carbonates than in mudstone, experimental studies show that
into clay-bound water, capillary-bound fluid siliciclastic rocks. For example, vugs (large the value of ‘A’ is 0.05 for kSDR and 100 for
and free fluid is the basis for many interpre- and often isolated pores) increase the ratio kBWP. These values were calibrated by MDT
tation methods which are discussed in the of pore body size to pore throat size permeabilities and conf irmed by core
following sections. (which for an isolated vug is zero!) but are permeabilities.
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 41
Searching for a correlation
Estimation formula RMS error Expected
exponent A comprehensive review of CMR-derived
permeability in Saudi Arabia examined
Estimation from T2 k = 0.13xf4T22 x/3.7 T22
results from five different reservoirs. The
k = 0.0073 (f2T2)3.3 x/3.3 T22 values of CMR-derived permeability and
porosity-derived permeability were com-
k = 1.3x10Ð9 T24.5 x/3.5 pared with core permeability values. The
Suncent25 = 3.2Ð1.2 log10 T2 +
Ð 0.09 T2Ð1 results are presented in Table 3.2, which lists
the correlation coefficients and residual
PHg = 1.6x106 T2Ð1.9 x/1.9 square errors for each method. The table
Estimation from porosity k= 2.7x105 f6.9 x/6.0 also shows the lithology-dependent cali-
bration constant ‘A’ used in the permeability
Wuncent25 = 110Ð2f +
Ð 18 equation and T2c, the bound water cutoff. In
Estimation from median k = 2x105 PHgÐ2.1 x/1.8 PHgÐ2
every case the correlation between NMR-
mercury pressure derived permeability and core permeability
is better than that between core per-
meability and porosity-derived permeability.
Table 3.1 Estimation formulae and estimation errors for T2. The symbol x/ indicates the average error factor in
the estimation. More specifically, it is the RMS error obtained in the fit of the logarithm of permeability (for
Figure 3.11 clearly illustrates the errors
example), translated back into the permeability domain that can occur when permeability is
calculated from porosity. In this example
the top panel displays permeabilities
Reservoirs Prefactor Predicted NMR permeabilty NMR permeabilty derived from three sources: cores, the
A average k= Af4 T22 log k= af+b
T2 cutoff k= f(f, Sw, Pc) Field C CMR tool and a fluid analysis log (FAL).
(msec) Conventional porosity/permeability cor-
Recent calibration relation assumes that permeability is a
Field A-X 0.05 295 corr=0.1, R=0.5 corr=0.02, R=0.4 monotonic function of porosity; however,
we have seen that this is not true in com-
Field A-Y 0.09 220 corr=0.12, R=0.5 corr=0.15, R=0.5 plex carbonates. This is illustrated by com-
Field A-Z 0.03 380 corr=0.4, R=0.5 corr=0.2, R=0.5 paring the two depths (A and B). At A the
porosity is 25%, and at B it is 28%, but core
Field B-XX A 1 66 corr=0.29, R=0.53 corr=0.26, R=0.51 permeability at A is three times that at B.
Field B-XX B 0.08 230 corr=0.34, R=0.53 corr=0.1, R=0.51 The trend of CMR-derived permeability is,
however, correct, and quite different to the
Field C-YY 1 0.01 660 corr=0.22, R=0.43 corr=0.21, R=0.3
FAL permeability derived from porosity.
Field C-YY 2 0.04 330 corr=0.3, R=0.5 corr=0.14, R=0.2 The lagoon facies (YY2) of Field C are
shown in Figure 3.12. The striking feature
Table 3.2 Correlation coefficients (corr) and residual square errors (R) for porosity-derived and of this reservoir is its almost constant
CMR-derived permeability estimates. The table also shows the lithology-dependent calibration porosity (25%) across the cored interval
constant ‘A’ used in the permeability equation and T2c, the bound water cutoff which comprises oil, transition and water
zones. While the porosity is almost con-
A porosity term such as ( f CMR ) 4 or The dynamic range of the T2 term may stant the core permeability varies from a
(fDN)4 is present in both relationships and is be attributed to a number of factors, such low of 2 mD to a high of 20 mD. The
interpreted to make the adjustments from as the presence of vuggy porosity. It may classical transform (which incorporates
pore body size to pore throat size. The also be due to bulk relaxation of water in saturation and capillary pressure terms)
functions of T2logmean or irreducible water large vugs or it may indicate that the ratio gives acceptable results in the oil zone
volume BVirr are related to pore size dis- of pore body to surface area does not scale (lower ellipse) but fails in the transition and
tribution and therefore permeability. with the pore throat diameter. These water zones. The trend in the water zone
Permeability has the dimensions of [L2] would of course invalidate the basic is the reverse of what would be expected.
and NMR T 2 units can be reduced to assumptions used to derive permeability. In contrast, CMR-derived permeability
length. Since porosity is dimensionless it This is the case in a peloidal micrite shows excellent agreement with the core
may be expected that T2 will appear raised sequence found in the Cretaceous permeability across the entire interval.
to the power 2 in expressions relating per- Thammama reservoirs of Mubarraz Field, The important conclusion to be drawn
meability and porosity. This is not necess- Abu Dhabi. The tangle of micrite crystals from this is that carbonate permeabilities
arily the case (Table 3.1) and selection of found at the surface of these peloidal grains derived from NMR T2 do not have the large
the appropriate relationship will be necess- produces a greater surface area than that dynamic range which would be predicted
ary depending on the lithology and per- for a ‘smooth’ spherical grain of the same from simple physical models. In practice this
meability of the reservoir rock. diameter. As the component crystals do not means that interpreters must be careful
N u m b e r 2 0 , 19 9 7
scale with the overall particle diameter they when choosing a model to derive carbonate
will alter the ratio of surface area to particle permeability from CMR measurements.
diameter (and pore throat diameter).
42 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
In carbonates the calibration of CMR 103
k core
Permeability (mD)
permeability must be done by individual
102 k CMR
facies or by reservoir units, due to their k Fal
diversity of pore size, connectivity and
101
mineralogy. However, as core/CMR per-
meability databases are populated with 100
increasing numbers of data points, from
laboratory measurements and from field 10Ð1
data, the calibration of CMR permeability 720 740 760 780 800 820 840 860 880 900
will improve. 50
A f core B
There is no doubt, however, that the 40 f CMR
Porosity (%)
permeability from the commonly used
NMR estimators is more reliable than from 30
porosity correlations alone. 20
10
Fast CMR logging
0
An important difference between the SDR 720 740 760 780 800 820 840 860 880 900
(ft)
and Timur relationships is their use of CMR
data. The SDR expression requires data Figure 3.11 A comparison of CMR and fluid analysis log methods for the derivation of
only from the CMR tool (total porosity and reservoir permeability
T 2 logarithmic mean), with permeability
being computed from the entire T 2 dis-
tribution without imposing a cutoff. The
Timur method, in contrast, uses only the
short T 2 data for B Virr evaluation, with
porosity being derived from the neutron- Porosity (%) Permeability (mD) k FAL = 5.67 f0.86
1000
density logs. This method does require a CMR CMR (Sw1.26 Pc0.89)
cutoff to be determined if BVirr is to be cor- FAL FAL 100
Oil k CMR (mD) corr = 0.3
rectly evaluated. Core Core
R = 0.2
One advantage of the Timur method
becomes apparent when only permeability 1100 Transition
and bound water volume (not total 10
porosity) are needed from the CMR Water
measurement. The period needed for
protons to align in the permanent magnetic
1200
field of the tool prior to spin tipping is k Core (mD)
1
greater for protons in large pores (due to 1 10 100
greater diffusion within these pores).
Protons in the pores containing irreducible k CMR = 0.04f4 T22
water align more rapidly. If only the 1300 100
k CMR (mD) corr = 0.3
relaxation of these protons is to be R = 0.5
measured the wait time of the tool can be
reduced to shorten each cycle and there-
fore increase logging speed. 1400 10
In this ‘fast’ logging mode the CMR tool
can acquire permeability data at speeds of
up to 3600 ft/hour. This is consistent with
the complementary density-neutron k Core (mD)
porosity data acquired by a combined 1500 1
0 10 20 30 40 1 10 100 1 10 100
PLATFORM EXPRESS* tool. In carbonates the
crossplot lithology is likely to be the most Figure 3.12 Logs from the lagoonal facies (YY2) in Field C. Porosity across the cored
reliable input for the Timur equation. interval (including oil, transition and water zones) is almost constant at 25%
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 43
The logging speed for the CMR tool is
intimately linked to the wait time, which Wait time = 0.38 sec
itself must be chosen to ensure measure- Logging speed = 1600 ft/hr
ment accuracy. Software is available to
Wait time = 4.5 sec
assist in setting suitable logging parameters Logging speed = 200 ft/hr
Signal distribution
(wait time, speed, ‘Te’ (echo spacing) and
number of echoes). This software models
relaxation time for formation fluids (oil and
water) based on their viscosity, oil density,
water salinity and temperature, and com-
putes the T1 polarization times needed to
ensure adequate proton alignment prior to
spin tipping. The wait time and logging
speed can be computed from these par- 33 msec
ameters (Figure 3.13). cutoff
Fast logging for bound fluid can also be
used to identify zones which will produce 3.0 T2 (msec) 3000
dry oil. In intervals where fSw from con-
ventional logs exceeds BVirr from the CMR Figure 3.13 Reducing the time allowed for proton polarization (wait time) results in incomplete polarization at
tool, it is likely that water will be produced. long T2 but has little effect on the short T2 times associated with bound water. This means that bound water
If fSw equals BVirr water-free production can data can be acquired at much greater speeds
be expected, regardless of the actual value
of the saturation.
Fluid pair Capillary pressure Interfacial tension
(psi) (dynes/cm)
Initial reservoir saturation and
capillary-bound water Mercury/mercury vapour 167 480
During accumulation in the reservoir non- Air/water 25 72
wetting oil will only enter those water-filled
pores with throats larger than a certain Oil/water 9 26
(critical) radius. This explains the presence
of capillary-bound water in the reservoir Table 3.3 Capillary pressures rescaled by the ratio of interfacial tensions
zone. This water will remain in place during
production while surrounding oil (in pores Water which remains ‘uncentrifuged’ must therefore give careful consideration to
with larger throats) is produced. Where all under a given capillary pressure equivalent selecting the correct value in each case.
of the water is present in micropores and all can be subjected to laboratory NMR Bimodal T2 distributions may be divided,
of the oil in macropores, dry oil can be pro- measurements to determine T 2 for the somewhat subjectively, into bound and free
duced, even when initial water saturations capillary-bound water. An alternative components ‘by eye’. The correct choice of
are extremely high. laboratory method to determine the T2c can subsequently be confirmed from
A measurement that identifies intervals capillary pressures necessary to remove observations of the height above the FWL
in which pore throats are smaller than the fluids is to measure the variation of water beyond which no water production occurs.
critical radius can indicate whether a zone saturation with mercury entry pressure.
will produce dry oil. The CMR tool’s T 2 The air/water or mercury/water press-
Variable T 2 cutoff
measurement can be used in this way to ures applied in the laboratory must be
define the amount of capillary-bound water. scaled to give equivalent reservoir capillary In some cases, for example where surface
pressures (Table 3.3). A 9 psi oil/water relaxivity is unknown and variable, a fixed
Laboratory methods capillary pressure corresponds, af ter value of T 2c gives misleading values for
scaling, to 25 psi air/water capillary press- capillary-bound water volume. The
Centrifuging a rock sample to remove water ure. This is equivalent to the capillary press- availability of core permeabilities provides a
from the pores is equivalent to a ure 60 ft above the FWL for oil and water means of determining T2c at each depth.
nonwetting fluid (air) displacing a wetting with specific gravities differing by 0.3. Figure 3.14 shows a carbonate in which
fluid (water) from the pore space under a The widely used T2 cutoff values (T2c) the bound fluid volume has been deter-
given capillary pressure determined by the for sandstones (33 msec) and carbonates mined using both a variable T2 cutoff and a
centrifuge speed. (100 msec) were obtained from numerous ‘default’ 100-msec carbonate cutoff. The
laboratory measurements at 100 psi air/ water volume from openhole logs, shown
water capillary pressure. The higher T2c in as a dotted line in track 3, exceeds the
carbonates is consistent with the lower bound water volume determined by the
values of surface relaxivity associated with default 100-msec cutoff. This would indicate
N u m b e r 2 0 , 19 9 7
calcite. It is likely, however, that these that there is producible water, but in fact,
values are general approximations that do the reservoir is a water-free oil producer.
not represent the capillary pressures The variable T2 cutoff correctly predicts
present in a particular reservoir. Engineers that there will be no water production.
44 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
Figure 3.14 In this
BFV at 100 msec
T2 Amp kND example the bound fluid
BFV variable T2 volume has been
L H kSDR determined using a fixed
kcore Free fluid (100 msec) and a
variable T2 cutoff. The
Producible water
1400
5000
water volume from
300
100 1000 100Ð1msec cutoff
14
64
0.1 1 10 openhole logs (the
3
800 fcore
900
fNMR
100 msec
T2 cutoff
Variable
T2 cutoff
Pseudo-capillary pressure curves Reservoir engineers often use lab-oratory Most descriptions of CMR logging
from CMR T 2 distribution methods to determine the pore space swept explain how examination of rock sections,
under a range of capillary pressures. or mercury injection experiments, can be
Capillary pressure curves plot f luid
Obtaining conventional capillary pressure used to demonstrate the equivalence
saturation versus applied (capillary) press-
curves by centrifuging or mercury injection between CMR T2 distribution and pore size
ure. They permit the residual water
tests on cores is time consuming and subject distribution. However, the same analysis
saturation expected at any capillary press-
to possible errors due to unrepresentative also shows that, for a particular pore
ure to be read off and are an essential tool
core samples. Fortunately, NMR measure- geometry, 1/T2 is proportional to Pc (the
for understanding the dynamics of reservoir
ments can be used to generate equivalent capillary pressure which would sweep pores
production. The diversity of pore sizes
pseudo-capillary pressure curves. to a certain saturation).
within carbonate reservoirs, and the
possibility of having oil-wet and water-wet
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 45
The CMR tool provides a way to record Figure 3.15 This
continuous pseudo-capillary pressure curves figure shows the
influence of pore
by applying suitable laboratory-derived cali- throats on
DvolumeHg L
bration to a plot of saturation versus 1/T2, S M laboratory- and log-
both derived from the T2 distribution. derived Pc. curves.
S M L NMR The top example
In practice, this is achieved by plotting Pressure
saturation (determined from integrating the shows three pore
T2 sizes – (S)mall,
area under the T2 curve) against T2. This (M)edium and
intermediate step then permits plotting of DvolumeHg (L)arge – all
L
1/T2 versus saturation, which is the pseudo- S M without throats,
capillary pressure curve. giving one-to-one
S M L NMR correspondence
The analysis assumes a certain relation- Pressure
between differential
ship between the pore throat diameter and pressure and T2
T2
the pore body diameter. This can be distribution. The
illustrated by mercury injection experi- DvolumeHg lower two
L examples show that
ments. When pore throat diameter equals S M
restricting the
pore body diameter there is a one-to-one
S M L NMR throat of a
correspondence between dif ferential Pressure
particular pore
pressure and T 2 . When the same pore shifts the equivalent
T2
throats are restricted there is a shift in the mercury result
mercury injection pressure (higher pressure compared to T2
needed to sweep a pore), with no cor-
responding shift in the CMR-derived T 2.
This is not surprising as NMR measure-
ments respond to the pore body, while the
mercury injection pressure responds to 3 msec 5000 0 .1 mD 1000 30 % 0 100
pore throat diameter (Figure 3.15). 734ft
Figure 3.16 shows a carbonate interval 80
that exhibits trimodal porosity. Track 1 con- 60
(psi)
tains the T2 distribution plot, track 2 is per- 40
meability derived from the CMR tool and
track 3 contains the density–neutron cross- 750 20
plot porosity, with the CMR-derived 100
porosity divided into small, medium and 788ft
kCMR fDÐN 80
large pore components.
Pseudo-capillary pressure curves com- 60
puted from CMR readings at three depths (psi)
800 40
are shown to the right of the log. The
pseudo-capillary pressure curve at 788 ft is 20
fCMR
as expected; the curves at 734 ft and 860 ft
end abruptly at low saturations. This is 100
860ft
because the CMR tool used to record the T2 80
distributions was unable to record the short- 850
60
est T2 values corresponding to the highest (psi)
capillary pressures. The validity of CMR- 40
derived Pc curves depends on recording the 20
shortest T2 times, which correspond to high
capillary pressures (Figure 3.17). 900 0 20 40 60 80 100
A comparison between Pc derived from Water saturation (%)
core and from the CMR tool shows reason- Small pores
T2
able agreement between the two methods Permeability Medium pores
distribution
over the ‘plateau’ zone at high saturations Large pores
which correspond to high T2 values. The
two results disagree at low saturations Figure 3.16 This carbonate interval exhibits trimodal porosity. Track 1 contains the T2
which correspond to low T2 values. distribution plot, track 2 is permeability derived from the CMR tool and track 3 shows
The most recent version of the CMR tool density–neutron crossplot porosity. The Pc values derived from core and those from the
has reduced the T 2 times which can be CMR tool are in agreement over the ‘plateau’ zone at high saturations which correspond
to high T2 values. However, the two results disagree at low saturations which correspond
N u m b e r 2 0 , 19 9 7
recorded from 3 msec to 0.3 msec, resulting to low T2 values. Note the excellent agreement between CMR total porosity and
in improved pseudo-capillary pressure curves. density–neutron porosity throughout the interval
46 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
Saturation in carbonates Even when microporosity is a small per- (msec) (%)
centage of total porosity it can cause large 1 0
The presence of capillary-bound water Log-derived Pc rises
errors in measured saturation, particularly in smoothly when short T2
within carbonate reservoirs makes it much
oil-wet carbonates. The resistivity path 10 (small pores) are 25
more difficult to evaluate hydrocarbon measured
through the microporosity acts in parallel
saturations from conventional resistivity
with the path through the large oil-filled Pc
measurements. 100 50
pores giving rise to low resistivity values.
The contribution which a particular T2
The resistivity of a reservoir with water
porosity type makes to rock conductivity is 1000 75
filled microporosity paths between con-
expressed in the value of the cementation
ductive layers (such as wackestones with
exponent ‘m’ in the familiar relationship Depth= 788ft
high microporosity) may be further reduced 10000 100
Ro = Rw/fm 0 20 40 60 80 100
by the ‘parallel circuit’ through the shoulder
Variations in ‘m’ are caused by changes (msec) (%)
beds. This may result in computed
in the geometry of the interconnected pore 1 0
saturations as high as 90% in zones which Log-derived Pc ends
space, which in this model is assumed to be
will flow no water during production. abruptly when no short
homogeneous throughout the volume 10 T2 (small pores) 25
This presence of water-wet micro-
under investigation. The existence of vugs, are measured
porosity in carbonates is in some ways
for example, will give rise to a large increase
analogous to the presence of dispersed clay 100 50
in the value of ‘m’. In a formation with 25% Pc
within clastic sediments. Shale-bound water T2
porosity, computed saturation changes by
provides similar low resistivity paths through
approximately 30% as ‘m’ varies from 1.7 1000 75
the rock that can lead to erroneous
to 2.1 – a range which is quite realistic for
saturation computations. Depth= 734ft
carbonate rocks. 10000 100
In shaly sandstone sequences the sands
Uncertainty in the ‘m’ value may be 0 20 40 60 80 100
are normally water-wet, so the shale pro-
resolved by using the EPT tool to provide a Water saturation (%)
vides a path through a water-wet oil
measurement of the invaded zone
reservoir. In these sediments the shale Figure 3.17 Correspondence between T2
saturation that is independent of the distribution (red) and derived Pc (yellow). The Pc
volume can usually be evaluated from con-
resistivity logs. When substituted into the curves are obtained by integration of all T2 times.
ventional logs but the distribution of the
Archie equation for invaded zone Each upper graph shows short T2 time (small pores),
shale, and its effectiveness as a conductivity hence a realistic Pc curve, while each lower graph
saturation, this additional measurement
path, is less easy to determine. However, does not show any short T2 time so the Pc curve
allows the interpreter to assess the value of
shaly sand saturation equations, such as the ends abruptly
‘m’ at each level.
Indonesia equation, have been developed to
Variable ‘m’ interpretation is a step in
solve for saturations in these cases.
the right direction when it comes to evalu-
ating saturations in carbonates, but it
Bound fluid Water T2 Threshold
assumes a uniform porosity texture
Possible free water Water Moved hydrocarbon 3 (msec) 3000
throughout the rock volume being Volume of water from RT Moved hydrocarbon Oil Porosity T2 Amplitude
investigated. 0.0 (%) 25.0 Oil 50.0 (%) 0.0 Calcite
More elaborate models have been devel- Perfs CMR bound fluid Diff. Caliper Dolomite
1:200ft 0.0 (%) 0.25 Ð20 (inches) 20 Anhydrite
oped to derive the ef fects of simple
X800
(spherical) vugs on the conductivity paths
through a conductive ‘host’ rock. These
models show that vugs contribute to the
conductive path, but rather inefficiently. The
CMR tool’s ability to differentiate large and
small vugs is a tremendous aid in applying
this model. Moved oil Residual
In some carbonates, vugs (macro- oil
porosity) are not the problem in determin- X900 Residual oil
ing saturation from resistivity. Instead the
presence of water bound in microporosity
Bound
may dominate the resistivity measurements, irreducible
giving saturations which are inconsistent water
with the production data (Figure 3.18). In
extreme cases this might mean producing
dry oil from an interval where Sw = 70%.
N u m b e r 2 0 , 19 9 7
Figure 3.18 The high water saturations recorded in some reservoir zones can be misleading. In this
example, conventional logs would suggest that water might flow from this zone. However, the CMR
tool shows that the water is bound in the rock’s micropores and that the zone will flow dry oil
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 47
Even when the proportion of micro- The basic approach is to assume the macro- CMR logs provide a new way to obtain
porosity in a dual porosity system can be and microporosity systems act in parallel, information about residual and heavy oils. In
evaluated with the same confidence as Vclay, each with its own ‘m’ and ‘n’ value. This water-wet rocks droplets of oil and gas lie
the solution of the saturation equation for approach must be modified to account for at the centre of the pore space and do not
the two situations is not entirely analogous. electrical interaction between the two contact the rock surface. They are, there-
Conduction through dispersed shale systems and for the fact that at least some fore, affected only by bulk relaxation and
involves cationic exchange, but in carbonate of the microporosity exists in dispersed not by pore properties or other formation
micropores it relies on ionic conduction. form and is contributing less to the total properties. In oil-wet rocks there will be sur-
Ionic conduction is the dominant process in conductivity than might be expected. face relaxation in addition to bulk
the macropores of both clastic sediments Obtaining the saturation of a carbonate relaxation, and in both cases relaxation due
and carbonates. with dual porosity is not easy, but at least to diffusion-in-gradient effects will occur.
A further difference is that in a clastic, part of the riddle is solved by knowing the Low viscosity fluids have large ‘diffusion
typically water-wet, sediment there is a con- component of the microporosity from CMR coef f icients’ which make them more
ductivity path along the surface of the data. As with the dispersed shale in a shaly susceptible to relaxation by diffusion in the
water-wet grains which acts in parallel to sandstone sequence this does not help to magnetic field gradient. This results in the
the shales even when hydrocarbon reveal the distribution of microporosity departure of the T2 relaxation time from its
saturations are high. This is not the case in throughout the rock but it is a first step expected relationship with viscosity at low
oil-wet carbonates where the only con- towards understanding its effect. The CMR viscosities. The relationship between bulk
ductivity path present will be through tool can measure microporosity directly, relaxation time (T2b) and oil viscosity at
water-wet micropores. making it possible to apply this model. reservoir conditions is shown in Figure 3.20.
For these situations interpretation The effects of diffusion are greatest at
methods have been developed based on long echo spacing and can be reduced by
Residual oil
dual porosity systems. Petrophysical par- using shorter Te (echo time) values while
ameters are defined for each porosity com- The ability to measure residual oil logging. At higher viscosities the T 2
ponent and ‘mixing laws’ are used to model saturation, capillary-bound water volume relaxation times are much shorter and the
the interaction between the two systems. and permeability is critical to evaluating the effects of relaxation due to diffusion are less
economics of carbonate reservoir develop- important. The relationship between T2 and
1.1 ment. Residual oil saturation is measured by viscosity is thus better in high viscosity oil.
logging tools in the invaded zone. Relationships between bulk relaxation
1.0 Many methods of evaluating residual oil time and oil viscosity have been developed.
saturation, including carbon/oxygen logs, For heavy crudes these methods yield T2
Hydrogen index
0.9 pulsed neutron logs and microresistivity values of about 5 msec. With heavy oils the
logs, have been used, but NMR logs provide predominant relaxation method is bulk
additional data. relaxation, and the nature of any contact
0.8
Residual oil saturation is also needed to between the oil and the grain surface is not
correct CMR-derived porosity. This porosity important. This permits viscous oils to be dif-
0.7
is free from lithology and environmental ferentiated from free fluids by a low T 2
effects, but the presence of heavy crudes, cutoff. The CMR-200 tool, which can read
0.6 with hydrogen index less than 1, influences short T2 times with improved accuracy, will
10 20 30 40 50
API gravity the measurement. The number of mobile permit better evaluation of heavy residual oil.
protons (Figure 3.19) is smaller in viscous oil Depending on the cutoff, however, the
Figure 3.19 Hydrogen index of crude oils as a function
than in water and these may not contact T2 distribution will include both capillary-
of their API gravity. Heavy crudes with a hydrogen
index below 1 influence CMR porosity measurements the grain surfaces in a water-wet sediment. bound water and heavy oil. The volume of
10
T1
T2 (Te = 0.32 msec) First pass prior to
1 doping: Vw+Vo is Water
T2 (Te = 1 msec) determined
T2b (sec)
0.1 Oil
T2
T2 (Te = 2 msec) Second pass after
doping. The water
0.01 is shifted below tool Oil
T2
detection threshold:
1/T2 = 1/T1 + (1/T2) D only Vo is determined,
0.001 ROS=Vo/(Vw+Vo)
0.1 1 10 100 1000 Detection threshold (3msec)
Viscosity (cP)
N u m b e r 2 0 , 19 9 7
48 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
viscous oil seen by the CMR measurement GR After doping Before doping ROS NMR stations Figure 3.22 In this
must be determined in order to establish T2 amplitude T2 amplitude
ROS Core k Residual example from a
CALI Low High Low High 0 1 (mD) Movable MIddle East well
corrections for CMR effective porosity and 4 9 5 (msec) 5000 5 (msec) 5000 1 SxoSw 0 5 0 0.3 f 0 residual oil
irreducible water volume. This may be saturation was
achieved by conventional saturation interpre- determined by three
tation in the filtrate-invaded zone using different methods.
resistivity devices such as the MicroSFL* tool The CMR results
compare favourably
or microlog devices to evaluate Rxo. The with cores in zones
absence of shales in carbonate reservoirs of good porosity
250
should simplify the interpretation. and permeability
Residual oil saturation can also be
fNMR
measured directly using techniques which after
suppress the T2 signal from the water by
doping the mud with manganese chloride
(MnCl2). The MnCl2 dramatically reduces 300 fNMR
the water’s T 2 such that it falls below before
either the 3-msec tool resolution or
another appropriate cutoff value. The fD
remaining signal is then attributed only to
residual oil in the invaded zone. The ratio
of oil-f illed porosity to total porosity 350
(determined from an initial CMR pass
before mud doping) yields residual oil
saturation (Figure 3.21).
This technique moves the water signal to
lower T 2 values, making it suitable for Cutoff Cutoff
measuring lighter oil in the invaded zone.
222ft 245ft
The method can optimized by setting the
‘echo time’ (Te) to maximize the contrast
between T2 for oil and T2 for water. Oil has
a much larger diffusion constant than water,
so a short echo time was used. The MnCl2
content is also adjusted to ensure an
appropriate T2 for the water.
To further refine the differentiation of
oil and water a T2 cutoff can be imposed on 1 10 102 103 104 1 10 102 103 104
the bimodal T2 distribution observed after (msec) (msec)
doping to eliminate any relaxation due to 256ft 274ft
imperfect doping.
Figure 3.22 compares residual oil
saturation determined by three different
methods in a Middle East well. The CMR
technique compares favourably with cores
in zones of good porosity and permeability.
Residual oil saturation determined by con-
ventional interpretation of openhole 1 10 102 103 104 1 10 102 103 104
resistivity compares poorly with both the (msec) (msec)
CMR and core results. 316ft 350ft
Tracks 1 and 2 in Figure 3.22 show
CMR-derived T2 distributions before and
after doping. Note the decrease in signal
amplitude due to the missing water signal.
Track 3 compares ROS from cores (red),
NMR ROS (black) and (1–Sxo) openhole
logs (blue on a matching scale). Track 4 dis-
plays NMR permeability and track 5 com-
pares density porosity with NMR total 1 10 102 103 104 1 10 102 103 104
(msec) (msec)
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 49
mismatch indicates either poor selection of (%)
matrix or fluid density (affecting fD) or the 100
Water T2 (msec)
presence of shale bound water in the total
Oil WT= 6 sec
porosity, and T2 falling below the CMR-200 80
tool’s threshold of 3 msec. (%)
60 W/o ~
Ð1
Typical T2 distributions after doping are 100
shown in Figure 3.23. The bimodal dis- 40
80 Water T2 (msec)
tribution is clear and selection of an Oil DSM 6 sec - 2sec
appropriate cutoff can be made without dif- 20
60
ficulty, in this case at 30 msec.
0
0 10 100 1000 10000 40
Hydrocarbon identification by the (%)
differential spectrum method 20
100
Water T2 (msec)
The difference in water and oil/gas polar- Oil WT= 2 sec 0
80
ization times is the basis for another 0 10 100 1000 10000
method for identifying hydrocarbons – the 60
differential spectrum method (DSM). Figure 3.24 Water and oil polarize at different rates
Short wait times result in incomplete T1 40 and therefore wait time affects the relative
polarization within the fluid. Water and oil amplitudes of the T2 relaxation signal due to the oil
polarize at different rates and therefore 20 and the water components. The spectrum of the
wait time affects the relative amplitudes of difference in T2 between the two passes shows peaks
0 corresponding to bulk relaxation times of water and
the T2 relaxation signal due to the oil and hydrocarbons
0 10 100 1000 10000
the water components. The spectrum of
the difference in T2 between the two passes
CALI
shows peaks corresponding to bulk 7 (IN) 12
2nd Repeat 1st Repeat Main Pass
0.4 s 6s 2.6 s
relaxation times of water and hydrocarbons
GR Rt NPHI
(Figure 3.24). 0 100
(GAPI) 0.1 1000 30% 0
In the method proposed by Shell in
Rxo DPHI
1200
1200
5000
1995, two T2 spectra are recorded with dif-
290
290
20
70
20
70
0.1 1000 30 % 0
5
1200
1200
290
290
290
20
70
20
70
20
70
Water
uncertainty of pore-size changes resulting in Water
a more robust hydrocarbon indication.
Figure 3.25 The transition from water to oil in this carbonate reservoir is clearly seen in the resistivity (track
4). The CMR logs, repeated for wait times of 6 sec, 2.6 sec and 0.4 sec, are shown in the top panel. The three
DSM spectrum differences are shown in the bottom panel (conventional DSM is shown in track 3)
50 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
Carbonate case studies Organic UBI CMR T2 AIT DEEP DPHI
SGR 0.1 (ohm-m)1000 NPHI
Experience in applying NMR techniques to CGR f CMR
k CMR
Middle East carbonate reservoirs is 0 (gAPI) 300 Low High 3 (msec) 3000 0.01 (mD) 100 40 (%) 0
increasing. This unique measurement is
being exploited in answering questions Bound
which remain unanswered by conventional Free
logs. The following NMR examples are
taken from around the region to illustrate 350
the applications of the CMR technique
described above.
Calcite
nodules
Kuwait
The Jurassic reservoirs of the Minagish and
Semi-opened
Umm-Gudair oil f ields in West Kuwait fracture
present difficulties in determining effective
porosity and permeability due to their com- Laminations
plex lithology. 400
The Najmah sequence (Upper Jurassic)
comprises interbedded carbonaceous lime- Fault?
stones and black shales. Primary per-
meability is often low but in some intervals
open fractures support high flow rates.
Figure 3.26 shows log data from a typical
Najmah well. The RHOB and NPHI logs
indicate apparent porosity of about 20%,
and deep resistivity is above 100 ohm-m,
resulting in Sw computing below 10%. The 450
only indication of why production tests yield
nothing is seen in the spectroscopic gamma
ray logs, CGR and SGR.
CGR (excluding radioactivity due to
uranium typical of organic deposits) is very
Figure 3. 26 The CMR tool reveals low effective porosity (track 4) and permeability
low throughout the interval. The high (track 3) in this composite log from the Najmah sequence. This explains the lack of
values of SGR indicate considerable organic production while testing a zone with high RHOB–NPHI porosity (track 4). Spectroscopic
material, which is consistent with the high gamma ray (SGR) shows a high uranium rich (organic) content when compared to CGR
resistivity values. (track 1), confirming that organic (bituminous) material caused the anomaly. The high
flow rates observed in the reservoir can be attributed to open fractures (track 2)
Examination of the CMR data indicates
that there is bound fluid in the invaded
zone. The very depressed T2 amplitude is
Organic k CMR DPHI Bound CMR T2
interpreted as low effective porosity and NPHI
SGR Free
low permeability. The organic material f CMR
CGR
remaining in the invaded zone is bitumen; 0 300 .01 100 .3 0 3 (msec) 3000
430
this results in the misleading RHOB–NPHI
and resistivity readings, which indicate oil
saturated porous reservoirs.
A second Najmah example (Figure 3.27)
shows interesting RHOB–NPHI behaviour 450
which can be resolved using the CMR tool.
At X452 ft the logs anticorrelate; below
this depth the logs correlate and show a
downward increasing porosity trend with
occasional peaks. This increasing porosity
470
trend is due to the increasing shale and
organic content apparent from the spec-
troscopic gamma ray curves.
The effective porosity is expected to
decrease downwards in the presence of shale
N u m b e r 2 0 , 19 9 7
490
and organic material, behaviour confirmed by
the CMR tool. The downward increase in Figure 3. 27 CMR T2 distribution confirms a reduction in pore size with increasing
shale content is supported by the decreasing depth. The porosity anomaly at X452 ft is due to a brecciated fault
pore size seen on the T2 distribution.
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 51
The feature at X452 ft is revealed by Organic UBI CMR T2 AIT DEEP DPHI
the UBI* Ultrasonic Borehole Imager and 0.1 1000 NPHI
SGR f CMR
DSI* Dipole Shear Sonic Imager logs as an
CGR k CMR
open fault containing breccia. The neutron 0 300 Low High 3 (msec) 3000 0.01 100 40 0
and CMR tools read the high porosity,
while the directional density log pad
happens in this case to have read the low
porosity of the breccia, resulting in 610
divergence of the porosity measurements at
this point. Below the fault lies a series of
porous units that degrade with depth into
Fractures
finer grained, less porous units. This is
clearly observed in the CMR pore size
630
decreasing with depth.
The Sargelu (Middle Jurassic) reservoir
sequence consists of densely cemented
packstones–wackestones and more
Figure 3.28 RHOB–NPHI porosity in the Sargelu reservoir sequence (track 5) is
argillaceous mudstones–wackestones, which affected by shaliness and dolomitization. Accurate CMR porosity allows determination
tend to be carbonaceous and pyritic. As in of more accurate water saturation for the whole formation. The permeability shown
the Najmah reservoir sequence, open by the CMR log at 623 ft (track 3) is due to a large open fracture (seen in the
fractures support high flows despite low borehole image in track 2)
primary permeability.
In the Sargelu sequence (Figure 3.28)
the porosity is typically low (less than 4% UBI DPHI Bound CMR T2 DSI Figure 3.29 CMR
from cores), and the upper intervals are NPHI Free Stoneley data complement
cleaner than the lower intervals. f CMR other logs. Here,
Low High 30 % 0 0.3 (msec) 3000 pore-size changes
Determination of porosity from
Pore size seen on the CMR log
RHOB–NPHI logs can be subject to large are not observed on
Micro Meso Mega
errors because of dolomitization and the UBI log, while
shaliness. The ability of the CMR tool to the fractures in the
measure effective porosity in this formation bottom interval,
590 seen by the UBI tool,
permits formation evaluation, even in the Pore size are shown, by the
shalier intervals, without resorting to shale distribution CMR log, to contain
corrections which are at best subjective cannot be bound fluid (blue in
The complex pore-size distribution seen from UBI track 2). These are
610 or DSI logs interpreted as being
which is encountered in the Sargelu
barely open
sequence cannot easily be inferred from
other logs. The UBI tool is sensitive to rock
texture but, as seen in Figure 3.29, it is not
able to distinguish the pore-size changes 630
seen by the CMR tool. The fracture visible
at the bottom of the interval on the UBI log
is shown, by the CMR log, to contain a high Semi-
bound fluid volume (blue in track 2) and is 650 opened
therefore interpreted as being barely open. fractures
This is confirmed from the absence of char-
acteristic chevron patterns produced by
open fractures on the DSI Stoneley VDL
presentation (track 4). It is clear that a
reliable interpretation cannot be obtained
Consequently prediction of zone pro- The estimation of formation per-
without the CMR data to complement
ductivity is impossible from conventional meability is of prime importance in this
other log data.
wireline log data. major reservoir. The expected trend of
Beneath the Najmah and Sargelu
Porosity in the Marrat Formation varies increasing permeability with increasing
reservoirs and beneath the Dharuma shales
from 0 to 15%: pore sizes range from porosity is seen in the interval 650–670 ft,
lies the Lower Jurassic Marrat Formation.
macropores to micropores. This complex pore but below 750 ft the CMR permeability
This clean limestone reservoir comprises a
structure is shown in track 2 of Figure 3.30 trend does not match the porosity trend.
series of grainstones and packstones, inter-
which uses T2 cutoffs from the T2 distribution The T 2 distribution shows that the per-
N u m b e r 2 0 , 19 9 7
52 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
The CMR permeability interpretation GR CMR T2 k CMR DPHI Bound
for this example is seen to correlate well k DSI NPHI Free
with permeabilities determined from the k RFT f CMR
0 GAPI 100 3 (msec) 3000 0.01 (mD) 100 30 (msec) 0
RFT* Repeat Formation Tester tool
mobilities (track 3). A Stoneley-derived per- 5
meability from the DSI log in this well is also
seen to match the permeabilities deter-
mined from the CMR tool. Pore size
Micro Meso Mega
Oman 650
The Lower–Middle Cretaceous limestones Microporosity Expected
of northern Oman are predominately a f-k trend
packstone to mudstone sequence. Porosity
measurements alone cannot predict these
variations in reservoir quality, making it dif-
f icult to identify the best zones. High
irreducible water saturation associated with
700
the wackestones in some of the structures
is believed to be the reason for the low
resistivities of 1.5 ohm-m in the hydro-
carbon zone. This unusual combination of
high water saturation and low resistivity Unexpected f-k trend
makes it difficult to identify zones of water-
Pore size Permeability increasing
free production and the depth of lowest oil increasing
production for reserve calculations, and to but porosity decreasing
750
determine the oil–water contact from con-
ventional logs.
The addition of a CMR tool and an MDT
tool dual packer module to the con-
ventional logs was introduced to determine
the volume of irreducible water and Figure 3.30 The correlation between porosity and permeability that is normally valid in clastic reservoirs can
thereby to predict water-cut potential. break down in carbonate reservoirs with complex pore geometry. In this example from the Marrat Formation in
Permeability from the CMR pore size a West Kuwait well the CMR pore size distribution (track 2) explains the actual relationship between porosity
and permeability. In the interval 750–765 ft porosity increases downwards, but the porosity comprises smaller
analysis was calibrated with the MDT dual pores (track 2) resulting in a decrease in permeability measured by the CMR, DSI and RFT tools (track 3)
packer drawdown permeability, confirmed
by core analysis, to provide a continuous
MDT Perm Figure 3.31
measurement, and has been subsequently
0.1 (mD) 1000 Continuous
validated with core data. In addition a CMR Core Perm (HE) permeability from
T2 distribution
fast logging pass was used in the horizontal Producible porosity both SDR and BWP
0.1 (mD) 1000 1 (msec) 3000
well interval to find bound fluid volume and k-SDR CMR porosity
equations is shown
permeability. This was necessary as logging 0.1 (mD) 1000 30 % 0 in track 1 with MDT
results. Core results
speeds on pipe conveyed tools cannot be confirm the
reduced sufficiently for full CMR T2 analysis continuous CMR
to be accurate. permeability
Integration of CMR, core and per-
meability measurements is illustrated in
Figure 3.31. CMR permeability was CMR
estimated using both the SDR and BWP
equations and calibrated with the MDT- MDT
derived mobility points. The CMR per-
meability was confirmed with the core
results six months after the petrophysical
evaluation.
The bulk volume water (BVW) was
determined from the AIT tool’s 1-ft log and
integrated with BVirr from the CMR log. The
CMR T2 cutoff of 100 msec for carbonates Core
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 53
Figure 3.32 illustrates the zones of T2 distribution Figure 3.32 By
water-cut potential. The MDT tool’s dual 1 (msec) 3000 AIT comparing BVirr
0.2 2.0 20 from the CMR log
packer module was used to obtain a (ohm-m) and BVW from
formation fluid sample. Over 100 litres of the AIT log, zones
water were pumped before oil shows were Window for of potential water
horizontal well cut can be
seen on the OFA* Optical Fluid Analyzer
module and a sample was taken. Production tested identified. Where
100 BOPD at BVW equals BVirr ,
This zone was production tested to con- as seen at the
30% water cut
firm the water-cut potential and produced position of the
100 BOPD with 30% water cut, thus ‘window’, zero
validating the evidence of potential water MDT dual packer water cut can be
module pumped expected
production from the comparison of the
104 litres water
CMR tool B Virr and bulk volume water
(BVW) from the AIT tool. The results were
used to select the horizontal well trajectory
in the upper section of the formation where
the permeability was good and the water-
cut potential was 0%.
The quality of fit between CMR and
core permeabilities in another well is seen in
Figure 3.33. In this example the per-
meability was determined from the CMR
tool using both the pore size (K-SDR) and
bulk volume irreducible (K-BWP) tech -
niques and is compared to the core-derived
permeability.
The BVW was determined from the AIT
90 inch resistivity and compared with the Figure 3.33 This
BVW AIT
Permeability example compares
bulk volume irreducible (B Virr) from the 0.3 0 CMR ‘fast BW
CMR tool. The blue shading illustrates that Core Perm T2 distribution pass’ permeability
the lower 5 m section will produce with Producible porosity with CMR ‘SDR’
0.1 (mD) 1000
water cut. Perm Fast BW pass DPHI and core results.
A final example from Oman shows the The fast pass
0.1 (mD) 1000 0.3 (ft3/ft3) 0
shows excellent
use of the CMR tool to determine the zone k-SDR BPV Fast pass T2 cut-off
agreement
of lowest oil production for reserve 0.1 (mD) 1000 0.3 (ft3/ft3) 0 1 (msec) 3000
calculations.
The long zone in Figure 3.34 which
appears to be a transition zone was tested
with the MDT dual packer module to
validate the CMR results. The formation k-SDR at 200 ft/hr
fluids were pumped with the MDT tool and BVirr
the OFA module was used to identify oil
and water potentials. The MDT test results
and the CMR log seem to agree on the CMR Perm
lowest producible oil. Below this depth at 1600 ft/hr
there is residual oil (seen on the resistivity BVW
curve), which is not movable or producible. Core
The CMR techniques described in the
article for analysing residual oil may be
applicable here.
These examples show that CMR pore
size derived permeability (K-SDR) from the
slow pass at 200 ft/hour and bulk volume
irreducible permeability technique (K-BWP)
from the fast pass at 1200 ft/hour compare
well with both MDT and core data as well
as with each other. This comparison illus-
N u m b e r 2 0 , 19 9 7
54 M i d d l e E a s t We l l E v a l u a t i o n Re v i e w
They also illustrate the CMR tool's MDT (OFA module) Figure 3.34 The depth
ability to predict a continuous permeability 50 litres at 20% of lowest oil
water cut production predicted
in the carbonates based on a single well cali- by the CMR log in this
bration to the MDT tool and subsequent example, was
validation with core data. confirmed by MDT
tests. The long
MDT (OFA module) transition zone below
The future 90 litres at 50%
MDT (OFA module) this depth reflects
water cut residual oil
The application of NMR logging to carbon- 226 litres at
Lowest 100% watercut
ate sequences is enhancing our under- producible
standing of these challenging reservoirs. oil
With the introduction of CMR-200 tools to
the Middle East advances in carbonate inter-
pretation will continue. ROS or
oil wet?
The CMR-200 tool responds to
extremely fast relaxation decays with T 2
times in the range 0.3–2 msec and its signal
processing increases the signal-to-noise ratio
by 70%, permitting faster logging.
These very fast decay rates correspond
to clay bound water, providing a new shale
indicator and accurate evaluation of shale
ef fects on porosity measurements.
100% water
Evaluation of shaly carbonates will therefore
benefit from the CMR-200 tool’s ability to
measure clay bound water at the lower end
of the T2 distribution.
N u m b e r 2 0 , 19 9 7
M i d d l e E a s t We l l E v a l u a t i o n Re v i e w 55