Estimation of dynamic petrophysical properties of water-bearing sands invaded with oil-base
mud from multi-physics borehole geophysical measurements
Zoya Heidari* and Carlos Torres-Verdín, The University of Texas at Austin
Summary
We develop and successfully test a new method to estimate
permeability, capillary pressure, and relative permeability
of water-bearing sands invaded with oil-base mud (OBM)
from multi-physics borehole geophysical measurements.
The inferred petrophysical properties of water-saturated
sands are used for calibration of equivalent properties in
hydrocarbon-bearing sands within the same sedimentary
sequence. Our estimation method simulates the process of
invasion between OBM and water. We iteratively adjust
porosity, permeability (mobility), capillary pressure, and
relative permeability in the simulation of invasion until
density, PEF, neutron, and resistivity logs are accurately
reproduced with numerical simulations from post-invasion
radial distributions of water saturation.
Examples of application include the cases of oil- and gas-
bearing reservoirs that exhibit a complete capillary fluid
transition between water at the bottom and hydrocarbon at
irreducible water saturation at the top. We show that the
estimated dynamic petrophysical properties in the water-
bearing portion of the reservoirs are in agreement with the
vertical variations of water saturation above the free water-
hydrocarbon contact, thereby validating our estimation
method. Furthermore, we show that the radial distribution
of water saturation inferred from resistivity and nuclear
logs can be used for fluid-substitution analysis of sonic
compressional and shear logs.
Introduction
Conventional well-log interpretation uses water-bearing
reservoir units to estimate connate-water resistivity and
rock-fluid properties. Recent publications describe the im-
portance of numerical simulation of well logs to estimate
static and dynamic petrophysical properties of rock forma-
tions (Malik et al., 2008; Salazar et al., 2007; Heidari et al.,
2009). The main challenge in all of these studies is the non-
uniqueness of the estimation. To circumvent this problem,
we develop a multi-physics approach based on the numeri-
cal simulation of gamma-ray (GR), photoelectric factor
(PEF), resistivity, neutron, and density logs, first in shale
intervals and next in fully water-saturated zones invaded
with OBM. The case of water-saturated sands invaded with
OBM is exceptional because the invasion process involves
non-miscible fluid displacement between OBM filtrate and
water wherein resistivity logs are only affected by water
saturation (salt mixing effects are non existent). Depending
on the dynamic petrophysical properties of the formation,
notably permeability (mobility), capillary pressure, and
relative permeability, the displacement of water with OBM-
filtrate will result in shallow, deep, sharp and/or spatially
smooth radial fronts of water saturation away from the
borehole wall. The corresponding effect on resistivity logs
with variable radial lengths of investigation will be either
“stacked” or “separated” logs. Because of the relatively low
values of formation electrical conductivity in this specific
case of analysis (water-bearing sands), induction resistivity
logs will exhibit enhanced sensitivity to radial variations of
water saturation away from the borehole wall.
Method
We begin the simulation process in a shale interval near the
water-saturated zone of interest. Shale intervals constitute
adequate depth zones to estimate properties such as shale
porosity, clay type, volumetric concentration of clay, and
bounded water resistivity. Estimation of such properties is
performed with the iterative numerical simulation of well
logs until securing an acceptable match between measured
and numerically simulated logs. After estimating shale
petrophysical properties we proceed with the analysis of
water-saturated sands to estimate static and dynamic petro-
physical properties.
The first step of the simulation process in water zones is
conventional well-log interpretation to calculate initial
petrophysical properties. Entry points for our simulation
method are average layer-by-layer petrophysical properties
obtained from either conventional petrophysical analysis of
well logs or available core data. The second step consists of
constructing rock-fluid properties and simulation of the
process of mud-filtrate invasion. Among the various avail-
able commercial simulators, we use GEM®, a composi-
tional reservoir simulator, for OBM invasion. Two-
dimensional (2D, radial and vertical, with implicit azi-
muthal symmetry) simulations are implemented in cylindri-
cal coordinates to model invasion in vertical wells under
the assumptions of cylindrical flow and permeability isot-
ropy. The petrophysical model is the input for fluid-flow
simulation; it consists of non-shale porosity, initial water
saturation, absolute permeability, rock-fluid properties
(relative permeability and capillary pressure curves), for-
mation fluid properties, invasion parameters, and formation
fluid and mud properties. We assume that rocks are water-
wet and that formation fluids are first-contact miscible. The
third step consists of numerically simulating well logs us-
ing as input the spatial distribution of fluid saturation re-
 Dynamic petrophysical properties of water-bearing sands

sulting from the simulation of invasion. We simulate GR, Field Applications

PEF, density and neutron logs using a fast linear iterative
refinement method (Mendoza et al., 2009). This final step
compares the simulated and measured logs to appraise their
agreement and consistency. It is found that the multi-
physics approach implemented to combine resistivity and
nuclear logs honors the intrinsic differences in volume of
investigation of the measurements involved, which leads to
reliable radial distribution of fluid saturation and consis-
tently reduces non-uniqueness in the estimation of dynamic
petrophysical properties. Iterations are intended to progres-
sively improve the agreement between measured and nu-
merically simulated resistivity and nuclear well logs by
making adjustments to the layer-by-layer static and dy-
namic petrophysical properties (Figure 1).
Figure 1: Iterative workflow adopted in this paper to estimate
unknown static and dynamic petrophysical properties in water-
bearing sands.
Three challenging field application examples verify the
efficiency and reliability of the proposed estimation
method. Drilling mud is OBM in all the three examples.
We chose the examples to encompass different ranges for
porosity and hydrocarbon density. The first field example
corresponds to a Paleocene siliciclastic sequence with a
sandstone dome located in the Central North Sea. The zone
under study includes oil with a gravity of 40° API, fol-
lowed by an active aquifer (Salazar et al., 2007). Average
non-shale porosity and permeability of the water-saturated
zone are 0.22 and 200 md, respectively. The second field
example is a Trinidad shaly-sand sequence located in a
prodelta sedimentary system. This formation contains gas
followed by a water-saturated zone. Average non-shale
porosity and permeability of the water-saturated zone are
0.14 and 500 md, respectively. The third field example is a
deepwater, Gulf-of-Mexico unconsolidated shaly-sand
sequence in a turbidite depositional system formed mainly
by channel levees. The sedimentary structure includes clay
laminations and massive intervals with moderate to good
grain sorting (Malik et al., 2008). Average non-shale poros-
ity and permeability of the water-saturated zone are 0.25
and 100 md, respectively. Figures 2a, 3a, and 4 show the
final match between numerically simulated and measured
array-induction resistivity and nuclear logs.

(a) (b)
Figure 2: (a) Comparison of numerically simulated (solid lines) and measured (dashed lines) gamma-ray logs (second panel from left), array-
induction apparent resistivity logs (third panel from left), and neutron and density logs (right-hand panel) for a water-saturated zone in a Central
North Sea gas field. Final petrophysical properties were estimated from the matching of field logs with numerical simulations. The left-hand
panel shows the spatial distribution of water saturation. Time of mud-filtrate invasion (tinv) is 2.5 days. (b) Comparison of the calculated water-oil
capillary pressure curve in the same field example with the vertical variation of water saturation rendered by the petrophysical interpretation of
well logs. The final capillary pressure curve was obtained from the matching of field logs with numerical simulations.
 Dynamic petrophysical properties of water-bearing sands

[b/e]
(a) (b)
Figure 3: (a) Comparison of numerically simulated (solid lines) and measured (dashed lines) gamma-ray logs (second panel from left), PEF logs
(third panel from left), array-induction apparent resistivity logs (forth panel from left), and neutron and density logs (right-hand panel) for the
water-saturated zone in a Trinidad shaly-sand gas field. Final petrophysical properties were estimated from the matching of field logs with
numerical simulations. The left-hand panel shows the spatial distribution (radial and vertical directions) of water saturation. Time of mud-filtrate
invasion (tinv) is three days. (b) Comparison of the calculated water-oil capillary pressure curve in the same field example with the vertical
variation of water saturation profile yielded by the petrophysical interpretation of well logs. The final capillary pressure curve was obtained from
the matching of field logs with numerical simulations.

To cross-validate our estimations, we compared the esti-

Figure 4: Comparison of numerically simulated (solid lines) and
measured (dashed lines) gamma-ray logs (second panel from left),
array-induction apparent resistivity logs (third panel from left), and
neutron and density logs (right-hand panel) across a water-
saturated zone in the deepwater Gulf-of-Mexico (oil field). Final
petrophysical properties were estimated from the matching of field
logs with numerical simulations. The left-hand panel shows the
spatial distribution (radial and vertical directions) of water
saturation. Time of mud-filtrate invasion (tinv) is three days.
mated capillary pressure curves in the first two field exam-
ples with the vertical water saturation profile calculated
from conventional petrophysical interpretation of well logs
(Figures 2b and 3b). This exercise indicates an agreement
in both trend and value between the water-oil capillary
pressure curve and the radial saturation profile within the
water-saturated sand invaded with OBM. In the third field
example, we observe a sudden separation between sonic
compressional- and shear-wave velocities in the water-
saturated zone. Figure 5 confirms that presence of light
hydrocarbon in that depth zone is responsible for such be-
havior. To quantify the sensitivity of well logs to static and
dynamic petrophysical properties, we constructed a syn-
thetic case based on the third field example. Figure 6 shows
the effect of rock porosity and permeability on the radial
invasion profile. Although the permeability-porosity ratio
remains constant, the radial length of invasion decreases
and the invasion front becomes piston-like with an increase
of both porosity and permeability. Moreover, this example
shows that the common assumption of piston-like invasion
in the presence of OBM is not valid in formations with low
porosity and permeability. The radially smooth variation of
water saturation affects well-log responses and conse-
quently conventional well-log interpretation. Figure 7 com-
pares array-induction resistivity responses for a range of
porosity and permeability values subject to a constant per-
 Dynamic petrophysical properties of water-bearing sands
meability-porosity ratio. The maximum separation between
resistivity logs takes place with low permeability and po-
rosity values, where the invasion profile is radially smooth,
thereby enabling the estimation of dynamic petrophysical
properties. The three field examples and the single-layer
synthetic case confirm that the proposed multi-physics
method enables the estimation of dynamic petrophysical
properties such as capillary pressure curves, relative per-
meability, and absolute permeability in the presence of
radially smooth invasion profiles.
Figure 5: Variations of compressional- and shear-wave velocities
due to variations of water saturation in a rock formation containing
light oil. Compressional velocity experiences a sudden increase
when water saturation is higher than 0.90. This effect gives rise to
a relatively large difference between compressional- and shear-
wave sonic velocities in the water-saturated interval.
Figure 6: Sensitivity analysis of the radial profile of water
saturation to permeability and porosity in a single-layer synthetic
case. Curves describe radial OBM mud-filtrate invasion fronts for
different rock types with varying porosity and permeability. The
permeability-porosity ratio is kept constant at 4[md]. Porosity
ranges from 0.05 to 0.35. Permeability varies between 20 and 140
[md]. Time of mud-filtrate invasion is three days in all cases.
Conclusions
We introduced a new method to estimate dynamic petro-
physical properties of water-bearing sands invaded with
OBM that synthesizes multi-physics borehole geophysical
measurements. The method used to combine resistivity and
nuclear logs quantitatively honors the differences in vol-
ume of investigation of the various measurements and con-
sistently reduces non-uniqueness in the estimation of dy-
namic petrophysical properties. Moreover, the estimation
method honors the physics of fluid flow in porous media
and takes advantage of the immiscibility of mud filtrate and
formation water. Cross-validation of the capillary pressure
curve with the estimated radial profile of water saturation
confirmed the accuracy and reliability of the method. We
also showed that the inferred radial profile of invasion en-
ables the numerical simulation of borehole sonic logs (P-
and shear-wave velocity logs) using fluid-substitution equa-
tions to match the corresponding measured logs. The accu-
racy of the proposed estimation method decreases in the
presence of shallow invasion and/or shock invasion fronts.
Acknowledgments
We thank BP, the Andrew Field Partner Group, and Kerr-
McGee (now Anadarko) for providing the data used in the
field studies. The work reported in this paper was funded
by The University of Texas at Austin’s Research Consor-
tium on Formation Evaluation, jointly sponsored by Ana-
darko, Aramco, Baker Hughes, BHP Billiton, BG, BP,
Chevron, ConocoPhillips, ENI, ExxonMobil, Halliburton,
Hess, Marathon, Mexican Institute for Petroleum, Nexen,
Pathfinder, Petrobras, RWE, Schlumberger, Statoil, TO-
TAL, and Weatherford.
Figure 7: Sensitivity analysis of induction apparent resistivity
measurements to permeability and porosity in the single-layer
synthetic case. Curves describe numerically simulated AIT
apparent resistivity curves for different radial lengths of
investigation. The permeability-porosity ratio is kept constant at
4[md]. Porosity ranges from 0.10 to 0.35. Permeability varies
between 40 and 140 [md]. Time of mud-filtrate invasion is three
days in all cases and OBM invades water-bearing sand.