You are on page 1of 11

SPE-172942-MS

Predicting Hydraulic Fracturing in a Carbonate Gas Reservoir in Abu Dhabi


Using 1D Mechanical Earth Model: Uncertainty and Constraints
Manhal Sirat, ADCO; Mujahed Ahmed and Xing Zhang, Schlumberger

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Middle East Unconventional Resources Conference and Exhibition held in Muscat, Oman, 26 –28 January 2015.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
In-situ stress state plays an important role in controlling fracture growth and containment in hydraulic
fracturing managements. It is evident that the mechanical properties, existing stress regime and the natural
fracture network of its reservoir rocks and the surrounding formations mainly control the geometry, size
and containments of produced hydraulic fractures. Furthermore, the three principal in situ stresses’ axes
swap directions and magnitudes at different depths giving rise to identifying different mechanical
bedrocks with corresponding stress regimes at different depths. Hence predicting the hydro-fractures can
be theoretically achieved once all the above data are available. This is particularly difficult in unconven-
tional and tight carbonate reservoirs, where heterogeneity and highly stress variation, in terms of
magnitude and orientation, are expected.
To optimize the field development plan (FDP) of a tight carbonate gas reservoir in Abu Dhabi, 1D
Mechanical Earth Models (MEMs), involving generating the three principal in-situ stresses’ profiles and
mechanical property characterization with depth, have been constructed for four vertical wells. The results
reveal the swap of stress magnitudes at different mechanical layers, which controls the dimension and
orientation of the produced hydro-fractures. Predicted containment of the Hydro-fractures within the
specific zones is likely with inevitable high uncertainty when the stress contrast between Sv, SHmax with
Shmin respectively as well as Young’s modulus and Poisson’s Ratio variations cannot be estimated
accurately.
The uncertainty associated with this analysis is mainly related to the lacking of the calibration of the
stress profiles of the 1D MEMs with minifrac and/or XLOT data, and both mechanical and elastic
properties with rock mechanic testing results. This study investigates the uncertainty in predicting
hydraulic fracture containment due to lacking such calibration, which highlights that a complete suite of
data, including calibration of 1D MEMs, is crucial in hydraulic fracture treatment.
Introduction
This study is part of FDP of a Carbonate Gas Reservoir in Abu Dhabi. This onshore carbonate gas field
that deposited in Lower Cretaceous Thamama group, consists mainly of limestone, and represents one of
the most environmentally sensitive areas. The marine environment is both highly sensitive and of great
ecological importance.
2 SPE-172942-MS

Figure 1—Work flow process for MEM construction and its subsequent use in various applications.

The objective of this study is to provide a geomechanical model of reservoir intervals in the
stratigraphic section of interest, primarily to assist in conceptual stimulation design and well engineering.
The main task is to develop Mechanical Earth Models (MEM) for four vertical wells.

1D Mechanical Earth Model Construction


The fundamental approach for characterizing mechanical layering is to integrate all available data or
indications of rock strength, deformability, in-situ stresses, and pore pressure into a Mechanical Earth
Model (MEM). This will be followed by calibrating this MEM as rigorously as possible using available
measurements and observations in order to arrive at an internally consistent representation of the key
geomechanical properties and parameters needed for subsequent analyses and engineering designs (Ali et
al., 2003; Plumb et al., 2000). A simplified workflow of MEM construction is illustrated in Figure 1 (for
details of the methodology please refer to Sirat et al., 2014 a&b).

Rock Mechanical Properties


The dynamic rock mechanical properties were derived using the sonic and density logs. While static rock
properties profiles were estimated using proprietary correlations published in general literature. Using
compressional and shear slowness logs data together with the bulk density log, dynamic elastic properties
were calculated using the equations below (Fjaer et al, 1992):
(1)

(2)

(3)
SPE-172942-MS 3

(4)

where ‘Gdyn’ is dynamic shear modulus in Mpsi, ‘Kdyn’ is dynamic bulk modulus in Mpsi, ‘Edyn’ is
dynamic Young’s modulus in Mpsi, ‘␷dyn’ is dynamic Poisson’s ratio, ‘␳b’ is bulk density in g/cm3, and
‘⌬ts’ and ‘⌬tc’ are shear slowness and compressional slowness (both in ␮s/ft), respectively.
Under the dynamic conditions involving propagation of an elastic wave, the rocks appear to exhibit a
stiffer response than they would under static loading such as in a mechanical laboratory test or under static
loading in the ground or near to a wellbore or completion. In fact, the weaker and more flexible the rock
is, the larger the differences between the dynamic properties and the static properties. Since wellbore
deformation and failure involves what is a relatively slow process compared to high frequency wave
propagation, static data are needed for geomechanics analyses.
No core test data related to static Young’s modulus was available from this field for the calibration
purpose. Correlations obtained from the lab test data in a nearby field were used to compare the estimated
static Young’s modulus. For all formations in selected wells, static Young’s modulus was estimated using
Modified Morales correlation using porosity (PHIT) and Edyn.
For intervals where porosity log was not available static Young’s modulus was estimated using Edyn
based on the lab correlation from the nearby field. It was noted that porous formations tends to have lower
static Young’s modulus compared to tight formations.
The estimated static Young’s modulus trend using Modified Morales correlation was found to be in
agreement with that obtained using correlations from the nearby field. Therefore, Modified Morales
correlation was used in this study.
In general the average estimated static Young’s modulus was found to be in the range of 5.0 – 8.0 Mpsi
in the reservoirs for the selected wells. Although the estimated static Young’s modulus trend in general
agrees with other correlations, but there seems to be significant difference in terms of the magnitude
across some intervals, this uncertainty will have a substantial effect on the stimulation design (hydraulic
fracturing). To reduce this uncertainty it is highly recommended to obtain data from lab test conducted on
core samples taken from the selected reservoir intervals and calibrate the MEM. Static Poisson’s ratio
values were assumed to be equivalent to dynamic Poisson’s ratio.
Unconfined compressive strength (UCS) was estimated using modified empirical equation (Chang
2004) given below, which correlates elastic moduli of the rock and UCS.
(5)

(6)

Where ‘UCS’ is the unconfined compressive strength (in MPa), whereas ‘Esta’ and ‘Gsta’ are dynamic
Young’s modulus and shear modulus (in GPa), respectively. The minimum value obtained from the two
equations was used to estimate the UCS. In general the average estimated UCS in the reservoirs was found
to be in the range of 9,500 – 10,500 psi for the selected wells.
Correlations obtained from the rock mechanics test data in the nearby field were used for comparison.
The estimated UCS trend using Chang’s correlation is in agreement with that obtained correlations from
the nearby field.
Friction angle (FANG) was estimated by applying correlation obtained from the rock mechanics test
data (THMM-ZONE-A) in the nearby field. The applied correlation is based on dynamic Young’s
modulus and is given below.
4 SPE-172942-MS

Figure 2—Elastic and strength properties for Well-1.

(7)

Where ‘FANG’ and ‘Edyn’ are the friction angle (degrees) and dynamic Young’s modulus (Mpsi),
respectively.
Tensile strength of the formation was used to evaluate the tensile failure of the borehole due to stress
concentration. Compared to UCS, tensile strength is very low, and is normally in the range of 1/12 to 1/8
of the UCS. We calculated the tensile strength as 1/10 of UCS in this study.
Figure 2 shows one example (Well-1) of the elastic and rock strength parameters for the selected wells
from this field in the selected zones of interest. The description of curves in these figures, from left to
right, is as follows:
● Track 1: Measured Depth (MD).
● Track 2: Formation lithology.
● Track 3: True Vertical Depth (TVD).
● Track 4: Compressional slowness (DTCO) and shear slowness (DTSM).
● Track 5: Bulk density log (RHOB) and total porosity (PHIT).
● Track 6: Dynamic Young’s modulus (E_dyn), static Young’s modulus (E_sta).
● Track 7: Dynamic (PR_dyn) and static (PR_sta) Poisson’s ratio.
● Track 8: Unconfined compressive strength (UCS) and tensile strength (TSTR).
● Track 9: Friction angle (FANG).
SPE-172942-MS 5

Figure 3—Vertical stress profiles for Well-1 to Well-4.

Vertical (Overburden) Stress


The overburden stress at a given depth point below ground surface is the weight of overlying earth
material and can be calculated by integrating a density log for the above formations given by the following
equation:
(8)

Where ␴v is the overburden stress, ␳ is formation density and g is gravity acceleration constant.
Unavailability of data for certain formations and quality of density logs can affect the calculated
overburden stress (Dyke, 1988). In the interval of missing density, density was extrapolated using a power
law. The curve can be described by the following equation:
(9)

Where ␳sur, A0 and ␣ are three fitting parameters, TVD is true vertical depth, WD (0 in this case) is
water depth and AG is air gap.
These parameters for the power law curve were determined by adjusting the three reference points
(shown as X in Figures 3) to match the exponential curve to the density log over the depth interval for
which density data was available. The vertical stress profile together with density logs of the four selected
wells are plotted in Figure 3. The density log coverage in all the selected wells was good in the zones of
interest. The estimated average vertical stress gradient was in the range of 1.00 to 1.05 psi/ft.
Horizontal Stresses
Several methods for identifying stress direction from wireline logs are available, including borehole
breakout orientation, natural and hydraulic fracture orientation, shear sonic anisotropy, and 3-component
6 SPE-172942-MS

vertical seismic profile. The orientation of a wide breakout is a good indication of the direction of
minimum horizontal stress in a vertical borehole as, for a vertical well (assuming principal stresses are
near-vertical and horizontal), maximum horizontal stress ␴Hmax is perpendicular to borehole breakouts
while the minimum horizontal stress ␴hmin is parallel to borehole breakouts.
No sonic or formation micro image (FMI) data was available for the selected wells to estimate the
orientation of the horizontal stresses. The orientation of minimum horizontal stress was assumed to be
135° (from North), based on the regional trend and plate tectonics of the Arabian Peninsula (Sirat et al.,
2012a and b; Johnson et al., 2005).
The poro-elastic horizontal strain model (Fjaer et al, 1992) was used to model the magnitudes of the
minimum (Equation 10) and maximum (Equation 11) horizontal stresses.
(10)

(11)

Here the two horizontal strains ␧x and ␧y may be compressional (i.e., for tectonic compression) or
extensional (i.e., negative, to represent lateral spreading), and can be treated simply as calibration factors
that can be adjusted to best-match the resulting stress estimates to any leak-off test data or specific modes
of rock failure seen in image logs. For the current study no data was available from FIT/LOT tests to
calibrate the minimum horizontal stress magnitude profile.
Therefore, the most rigorous process available for estimating the complete state of in-situ stress in the
underground (i.e., magnitudes, order and directions) at a single well location, and particularly to determine
the magnitude of maximum horizontal stress ␴Hmax, involves some initial estimate of the range of possible
stresses. This is constrained by limit state or other mechanical considerations, that is then further
constrained and refined to generate a more specific prediction. This is further calibrated and validated by
increasing degrees, using as much information as available, to achieve a final model that is internally
consistent with and calibrated to all available data and stress indicators.
The horizontal stress magnitudes were modeled using Equations 10 and 11, and constrained using the
computed rock elastic properties generated earlier during the MEM construction.
Based on Andersonian classification of faults and their relationship to the stress regimes (1951), the
stress regime down to NAHR UMR Formation is predominantly normal faulting while within the
Formations below stress regimes vary between normal to strike slip faulting. Fine tuning of the models
was achieved by adjusting the strain terms in the equations for horizontal stress, to best match the
occurrences of breakout seen on caliper logs in the study wells. Figure 4 shows the profiles of the
minimum and maximum horizontal stresses for the selected wells. The average minimum horizontal stress
gradient was in the range of 0.84~0.88 psi/ft while the average maximum stress gradient was in the range
of 0.92~1.01 psi/ft in THAMAMA formation. The alternating overlapping between the vertical stress (␴v;
in red) and the maximum horizontal stress (␴Hmax; in blue) can define the swap of different stress regimes
with depth (Talbot and Sirat, 2001), such that when ␴v⬎␴Hmax it is a normal faulting, while (␴Hmax ⬎␴v
it is a strike-slip stress regime.
Identification of Potential Hydro-Fracture Zones
The propagation, orientation and shape of hydro-fractures are controlled by many factors, including in-situ
stress state, mechanical properties and operational factors. In this study, only geomechanical related
factors, such as in-situ stress state and mechanical properties were considered. To identify the potential
hydro-fracture zones in the field, the following criterions were adapted:
● Low fracture gradient for hydro-fracture zones
● High stress contrast between hydro-fracture zones and barriers
SPE-172942-MS 7

Figure 4 —Horizontal stress profiles for Well-1 to Well-4.

● Low YM for hydro-fracture zones and high YM for barriers.


Based on the 1D MEMs developed for the selected wells from this field, potential zones for fracturing
the selected reservoirs in THAMAMA formation are shown in Figure 5. The highlighted zones have lower
minimum horizontal stress; hence hydraulic fractures may initiate at these intervals. The zones of higher
stress above/below these intervals may act as a barrier depending upon hydraulic fracturing job size and
pumping pressure constraints.
Uncertainty in Mechanical Earth Model
An accurate geomechanical earth model (GEM) including constraints or uncertainty on stress magnitudes
and orientations, and mechanical rock properties are essential to understanding reservoir response for
stimulation and production, particularly in low permeability reservoirs.
In this study different correlations were used to derive the Young’s modulus profiles along the well
trajectories, including correlations based on 1) dynamic Young’s modulus, 2) modified Morales (poros-
ity), 3) dynamic Young’s modulus from the nearby field, and 4) porosity from the nearby field. The rock
strength profile was also derived using different correlations, including 1) Chang’s correlation, 2) porosity
based from the nearby field and 3) dynamic Young’s modulus based from the nearby field.
Due to lack of laboratory rock mechanics testing, the mechanical properties, both elastic and strength
parameters could not be calibrated. Therefore, calibration is necessary to reduce such uncertainty in
mechanical property characterization. This can be achieved by performing necessary laboratory tests,
including single/multi-stage tri-axial tests, unconfined compressive strength (UCS) tests, ultrasonic
velocity (UV) test (Vp/Vs) etc., to obtain rock strength and failure parameters more accurately. The data
from these lab tests will help minimize the uncertainties in estimated rock properties from the model and
further assist in calibrating/refining 1D MEM parameters.
8 SPE-172942-MS

Figure 5—Potential zones for the hydraulic fracturing in the reservoirs which are marked with the rectangular bars (upper: above THMM-D, lower:
below THMM-D).
SPE-172942-MS 9

In general, rocks with higher elastic modulus exhibits higher horizontal stresses. However, the
uncertainty in the estimated minimum horizontal stress magnitude is also significant, particularly in the
maximum horizontal stress. The uncertainty can be reduced by calibrating the derived minimum hori-
zontal stress magnitude profiles with any measurements, such as extended Leak-off test (ELOT),
mini-Frac test and formation integrity test (FIT), if available.
The orientations and size of produced hydro-fractures are largely governed by the in-situ stress regimes
of the reservoir rocks and the surrounding formations. Hence the hydro-fractures are mainly horizontal to
sub horizontal under the thrust stress regime, vertical to sub-vertical under the normal faulting or
strike-slip stress regime, or both. Furthermore, the three principal in situ stresses’ axes may swap
directions and magnitudes at different depths giving rise to identifying different mechanical bedrocks with
corresponding stress regimes at different depths. This is particularly true in tight carbonate reservoirs
where reservoir heterogeneity and highly stress variation, in terms of magnitude and orientation, are
expected.
Based on the analysis, the stress regime is constrained to be predominantly under normal faulting
(␴v⬎␴H⬎␴h) till NAHR UMAR while in formations below NAHR UMR till HABSHAN varying stress
regime from Normal to Strike-slip (␴H⬎␴v⬎␴h) was observed.
Due to the alternation in stress regime, the shape of the induced hydro-fractures is likely to change.
Under a strike slip stress regime, a low and long fracture is likely to create, but under a normal faulting
stress regime, a high and short fracture is likely to create. In this case, the risk to break the barriers is
higher. Therefore, it is important to understand the uncertainty in stress regime at those depth intervals of
interest. Because the uncertainty in stress regime is related to the maximum horizontal versus vertical ratio
stress, hence it can be reduced once the uncertainty in the maximum horizontal stress magnitude is
reduced.
No image, multi-arm caliper or sonic scanner data was available to estimate the direction of the
horizontal stresses. Based on the regional stress map and the measurements in the nearby fields, the
maximum horizontal stress direction of 45° (from North) was adopted.
As the propagation direction of the vertical or sub-vertical hydro-fractures is in the maximum
horizontal stress direction under a normal faulting or strike slip stress regime, the uncertainty in the in-situ
horizontal stress direction has an important impact on the hydro-fracture direction.
The control in the direction of hydro-fractures is crucial during reservoir stimulation treatment because
productivity is strongly related to the induced hydro-fracture direction. This is particularly true for tight
reservoirs where natural fractures usually can be utilized to provide additional flow channels. In this case,
it is desirable to intercept the natural fractures by the induced hydro-fractures for better communications
between producers and the pay zones.
The constraint in horizontal stress direction is also important in determining the horizontal well
azimuth for multistage fracturing along horizontal well. The maximum contact area can be achieved with
multistage transverse hydro-fractures along horizontal wells if the direction of the horizontal wells is
orientated in the minimum horizontal stress direction.
Discussions and Conclusion
This study suggests that geomechanical study is essential for optimal stimulation design in the FDP of this
carbonate gas reservoir. The 1D Mechanical Earth Models (MEMs) constructed can provide valuable
information in the understanding of the three principal in-situ stresses’ profiles and mechanical property
characterization.
The characterization of the mechanical layering resulting from the 1D MEMs was used to identify the
potential hydro-fracture zones.
The stress regime in this field is predominantly normal faulting (␴v⬎␴H⬎␴h) till Nahr Umr while
below Nahr Umr till Habshan varying stress regime from Normal to Strike-slip (␴H⬎␴v⬎␴h) is observed.
10 SPE-172942-MS

The understanding of stress regime alternation provides insight to predict the hydro-fracture shapes within
the reservoir intervals.
Based on the values of the mechanical properties, such as Young’s modulus and strength properties,
and the magnitude and orientation of the three principal stresses profiles, the parameters of the hydro-
fracture treatment can be optimized.
However, due to high uncertainty in 1D-MEMs constructed, resulting from lack of calibration of the
stress magnitude and orientation as well as mechanical properties, it is recommended to acquire a
complete suite of data, which is crucial in hydro-fracture treatment. The recommended data acquirements
include:
● Minifrac and/or LOT with FMI before and after to calibrate the magnitude of the minimum
horizontal stress
● FMI image and multi-arm caliper logs to calibrate the direction of in-situ horizontal stresses
● Interpreted sonic fast-shear azimuth to understand the magnitude and direction of in-situ horizontal
stresses
● Core sample rock mechanics tests to calibrate the mechanical properties of 1D MEMs.

Acknowledgements
The authors would like to thank ADCO and Schlumberger for their support and permission to publish this
work.

Glossary
␣ ⫽ Poroelastic coefficient
v, PR ⫽ Poisson’s Ratio
␳, ␳b, RHOB ⫽ Density (bulk), g/cm3
␴h ⫽ Minimum horizontal stress, psi
␴H ⫽ Maximum horizontal stress, psi
␴v ⫽ Vertical (overburden) stress, psi
␧x,y ⫽ Cartesian horizontal strains
⌬tp, ⌬ts ⫽ Compressional/Shear Sonic slowness, us/ft
AG ⫽ Air Gap, ft
DTCO ⫽ Compression sonic slowness, ␮s/ft
DTSM ⫽ Shear sonic slowness, ␮s/ft
dyn ⫽ Dynamic
E, YM ⫽ Young’s Modulus, Mpsi
ELOT ⫽ Extended leak-off test
FANG ⫽ Friction Angle, deg
FDP ⫽ Field Development Program
FIT ⫽ Formation Integrity Test
FMI ⫽ Formation Micro Image
G ⫽ Shear Modulus, Mpsi
g ⫽ Gravitational acceleration constant, m/s2
GR ⫽ Gamma Ray, gAPI
GEM ⫽ Geomechanical Earth Model
K ⫽ Bulk Modulus, Mpsi
LOT ⫽ Leak-off test
MEM ⫽ Mechanical Earth Model
MD ⫽ Measured Depth, ft
SPE-172942-MS 11

Pp ⫽ Pore pressure, psi


PHIT ⫽ Total Porosity, v/v
sta ⫽ Static
TSTR ⫽ Tensile Strength, psi
TVD ⫽ True Vertical Depth, ft
UCS ⫽ Unconfined compressive strength, psi
UV ⫽ Ultrasonic Velocities, m/s
Vp, Vs ⫽ Primary and Secondary sonic velocities, m/s
WD ⫽ Water Depth, ft

References
Ali A. H. A., Brown T., Delgado R., Lee D., Plumb D., Smirnov N., Marsden J. R., Prado-Velarde
E., Ramsay L., Spooner D., Stone T., & Stouffer T., (2003). Watching Rocks Change - Mechanical
Earth Modeling; Oilfield Review; Summer 2003, Volume 15, Number 1, pp. 22–39.
Anderson, E.M. 1951. The Dynamics of Faulting. Oliver and Boyed, Edinburgh, 206 p.
Chang, C., (2004). Empirical Rock Strength Logging in Boreholes Penetrating Sedimentary Forma-
tions. Geol. Environ. Science, Chungnam National University, Daejeon. Vol. 7, No. 3, pp
174 –183.
Dyke C. G., (1988). In-situ Stress Indicators for Rock at Greater Depth; Ph.D. Thesis, Imperial
College, University of London.
Fjaer E., Holt R. M., Horsrud P., Raaen A. M., & Risnes R., (1992). Petroleum Related Rock
Mechanics Developments in Petroleum Science, 33; Elsevier.
Johnson, C.A., T. Hauge, S. Al-Menhali, S. Bin Sumaidaa and B. Sabin, B. West, (2005). Structural
Styles and Tectonic Evolution of Onshore and Offshore Abu Dhabi, UAE. IPTC 10646.
Plumb, R. A., (1994). Influence of Composition and Texture on the Failure Properties of Clastic
Rocks. SPE 28022.
Plumb, R. A., Edwards, S., Pidcock, G. & Lee, D., (2000). The Mechanical Earth Model Concept and
its Application to High-Risk Well Construction Projects. SPE 59128.
Sirat, M., Zhang, X., Simon, J., Vantala, A., Povstyanova, M. (2014a). Mechanical Layering:
Implication for Hydraulic Fracturing in Unconventional Plays in a Tight Carbonate Field in Abu
Dhabi, UAE. SPE 172942, p. 1–9.
Sirat, M., M., Xi, G., Ni, Q., Mohamad-Hussein, A. and Zhang, X. (2014b). Methodology and
Implications for Natural Fracture Network Modeling in a Carbonate Reservoir of Abu Dhabi.
IPTC 17593, p. 1–14.
Sirat, M.; H. Koyi; S.L. Mahmoud and D.-M. Popa, (2012a). Conceptual Structural Model of an
Onshore Carbonate Hydrocarbon Field in Abu Dhabi: Constraints and Implication. SPE-161544,
P. 1–10.
Sirat, M. T. Al-Dayyani, M. Singh A.S. Abdul Rehman N.O. Al-aabi S. Mahmoud M. Shuaib and M.
Moge, (2012b). Characterization and Modelingof Fractures in arge Onshore Abu Dhabi Car-
bonate Field. SPE-162342, P. 1–14.
Talbot, C.J. and Sirat, M., (2001). Stress Control of Hydraulic Conductivity in Fracture-Saturated
Swedish Bedrock. Engineering Geology, 61, pp. 145–153.

You might also like