SPE 138869

3D Geomechanical Modeling for the Apiay and Suria Oil Fields

(Llanos Orientales Basin, Colombia): Insights on the Stability of Reservoir-
Bounding Faults
F. Fernández-Ibáñez, GeoMechanics International; F. Cardona, Ecopetrol; M. Holland, E. Araujo, M. Brudy, and
J. Alvarellos, GeoMechanics International; Y. Ordóñez, Ecopetrol; and D. Mateus, Instituto Colombiano del Petróleo

Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Latin American & Caribbean Petroleum Engineering Conference held in Lima, Peru, 1–3 December 2010.

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
Understanding the spatial variation of stresses and rock mechanical properties is the key to optimize strategies for field
development. In geologically complex regions, as faulted and folded domains or areas of lateral facies changes, integration of the
geologic structure is crucial to better predict geomechanical attributes. This has been done in an area located in the Llanos
Orientales Basin (Colombia) by means of a three-dimensional geomechanical model. The model integrates knowledge of the
geological structure and regional stratigraphy, data collected in more than 40 wells, and the drilling experience acquired over 20
years. The result is a high resolution 3D model, easy to handle and to update, with a number of immediate applications that go
from field exploration to harvesting and abandoning.
Seismically resolvable faults and horizons have been brought into three-dimensions to build a structural model of the Apiay-
Suria region. The structural framework has been used to propagate the geomechanical attributes of a number of wells using
geostatistical algorithms. As a result, we have built a 3D model comprising the magnitudes and orientations of the three principal
stresses, the pore pressure regime, and rock mechanical properties as uniaxial rock strength, internal friction or Poisson’s ratio.
The resulting 3D model provides a high resolution database that maps any lateral variation of the geomechanical attributes
across the field and can be used to assess wellbore stability for any future well trajectory and location. The 3D numerical
geomechanical model has been also used to determine the stress conditions near the main faults. The stress tensor is here combined
with a failure criterion to assess the potential for fault instability. This study provides a valuable insight on the risk of losses while
drilling through these faults and the possibility of seal rupture and leakage during injection and hydraulic fracturing activities.

Introduction
Development of most of Colombian oil basins is tightly related to the evolution of the convergent margin of western South
America where the Nazca, Caribbean and South American plates interact. In Colombia, the Andes form three distinctive orogenic
belts: the Eastern, Central and Western Cordilleras. The Eastern Cordillera is bounded to the east by a large Neogene and less
deformed foreland basin (Bayona et al., 2008), the Llanos basin, where the Suria and Apiay oilfields are located.
The tectono-stratigraphic units distinguished in the study area has been extensively described (e.g., Schamel, 1991; Gomez et
al., 2005; Parra et al., 2009, among others) and they are the following (from bottom to top): (1) Une Formation, (2) Gachetá
Formation, (some authors group Une and Gachetá under a common name: Guadalupe Group, (3) Mirador Formation; (4)
Carbonera Formation; (5) León Formation, and (6) Guayabo Formation. From an operational point of view, these formations are
divided in units and assigned a different nomenclature (see ages and correspondences between stratigraphic and operational units
in Table 1).
The Apiay-Suria region is affected by a system of Mesozoic normal faults with north and west vergencies. Some of these
faults were reactivated as reverse faults during basin inversion (e.g., Sarmiento-Rojas et al., 2006). A later stage of wrench
tectonics that lasts up to recent times developed a new set of strike-slip faults and reactivated some old ones. The western domain
of the Llanos basin is thought to be influenced by the large transpessional setting that develops along the Andes foothill
2 SPE 138869

(Colmenares and Zoback, 2003). Focal mechanisms show the coexistence of strike-slip solutions with strike-slip to reverse faulting
focal solutions depicting a complex pattern of stresses across the Llanos basin.
Wellbore image logs, reservoir pressures, wireline data, and drilling experiences from more than 40 exploration and
production wells have been reviewed to develop constraints on the in situ stresses and the rock strength in the Apiay-Suria region.
Since these parameters can vary spatially, both locally within a structure, and regionally within the basin, a 3D geomechanical
model has been developed to capture any variation across the field. This numerical model has a number of applications amongst
which wellbore stability and fault seal analyses are the most relevant. The advantages of three-dimensional geomechanical models
are extensively discussed by Holland et al. (2010).
The aim of this work is to use the knowledge of the present-day stress tensor to characterize the current state of stress on the
main reservoir-bounding faults across the Apiay-Suria region. The results are expected to provide insights on how to design and
minimize risks related to fault reactivation during hydraulic fracturing and water injection operations in the Apiay and Suria
oilfields. Hydraulic fracture stimulation will be performed during well completion or work-over to clean wells and enhance field
productivity. Unfortunately, such techniques may also carry additional operational risks such as casing damage from natural
fracture and fault instability or hydrocarbon leakage. Predicting the manner in which hydraulic fractures are likely to propagate
and the potential impact of this propagation is critical to the fracture stimulation design process.

Age Nomenclature
Era Serie / Stage Stratigraphic Operational
Miocene/ Pliocene Guayabo
Miocene León
Upper Carbonera SST
E Shale
C1
Cenozoic

Carbonera

Oligocene/ Carbonera SST

Miocene C2
E3 Shale
T1
E4 Shale
Eocene Mirador T2

Gachetá K1
Mesozoic

Cenomanian/
Maastrichtian
Une K2

Table 1 - Age and correspondence between stratigraphic

and operational units.

Fundamentals
Fault mechanics
Shear failure under compressive stress states is commonly described with the Coulomb-Mohr criterion. According to this criterion,
Coulomb frictional failure of a cohesionless fault will occur if τ, the shear stress, and σn, the effective normal stress (σn = Sn – Pp,
where Sn is the normal stress), are related by
τ = μσn (1)
where μ is the coefficient of sliding friction. The shear stress and effective normal stress acting on each fracture plane are
computed using the principal stress magnitudes, the fluid pressure, and the orientation of the fracture plane with respect to the
orientations of the principal stresses (see Jaeger and Cook, 1979).
SPE 138869 3

The Coulomb Failure Function (CFF) can be defined as,

CFF = τ − μ(Sn − Pp) (2)
The Coulomb failure criterion can also be solved in terms of the pore pressure (called the critical pore pressure, Ppcrit) at
which a fault element will begin to slip by setting CFF=0.
Ppcrit = Sn – τ/μ (3)
Ppcrit and CFF are usually computed using a coefficient of sliding friction of 0.6 based on laboratory measurements of the
frictional strength of prefractured rock (Byerlee, 1978). The difference between the critical pore pressure and the reference pore
pressure is called the critical injection pressure (ΔPpcrit = Ppcrit –Pp) (Fig. 1). ΔPpcrit is a measure of excess pressure above the pore
pressure that each segment of the fault can withstand before it slips. It is a measure of the probability for fault-related fluid flow.

Fig.1 - The Mohr Circle. Stresses acting on a planar feature can be plotted in the Morh space in terms of the normal
(Sn) and shear (τ) stresses acting on it.

The Mohr diagram is a tool commonly used to present the relationship between the normal (Sn) and the shear (τ) stresses
acting on a particular surface (Fig. 1). S1 and S3 represent the maximum a minimum principal tectonic stresses acting on a specific
fault surface. When pore pressure (Pp) is considered, the magnitudes of the principal stresses are reduced, resulting in what is
known as effective stresses (σ1 and σ3). Fault strength or sliding friction (μ) is an intrinsic property of each fault or fracture and is
used as an approach to how strong a fault is. This property is usually plotted together with the Mohr circles as a line whose slope is
defined by fault strength (the lower the slope, the weaker the fault, and therefore, the more likely to slip).
If the ratio of the shear stress to effective normal stress exceeds the sliding friction, the probability of shear failure occurring
is high, and then, the fracture is said to be critically-stressed (Fig. 1). If shear failure occurs, the seal membrane within the fracture
will fail and fracture permeability will form. It has been demonstrated that fractures with high ratios of shear stress to normal stress
tended to be hydraulically conductive, while fractures with lower ratios of shear stress to normal stress tended not to be
hydraulically conductive. Statistically, hydraulically conductive fractures are more likely to be critically-stressed (Barton et al.,
1995).
Constraining the In-Situ Stress Tensor
The stress tensor consists of a good knowledge of pore pressure, and in-situ orientation and magnitudes of the three principal
stresses (the vertical stress, Sv; the maximum horizontal stress, SHmax; and the minimum horizontal stress, Shmin). Following
Anderson (1951), we assume that one of the three principal stresses must be vertical, because the horizontal Earth’s surface must
be shear stress free. Following the strategy described by Zoback (2007), the field-wide state of stress was estimated using the
techniques described hereafter.
• The vertical stress (Sv) is estimated by integrating density distribution with depth. Overburden profiles are usually
generated after filtering density logs to remove spurious data due to poor hole conditions. In these situations, an alternative
method to derive the density profile is to use the sonic log to construct a pseudo-density curve (Gardner et al., 1974).
4 SPE 138869

• The magnitude of the least principal stress (S3) is commonly determined using hydraulic fracturing techniques (e.g. leak-off
tests) whereby the leak-off pressure is equated to the minimum principal stress. Although leak-off pressures do not yield as
a reliable estimate of the minimum horizontal stress magnitude; it is widely accepted that the lower bound to leak-off
pressures in vertical wells gives a reasonable estimate of the minimum horizontal stress (e.g. Breckels and van Eekelen,
1982).
• The pore pressure gradient in clastic sedimentary sequences is typically determined using wireline logs (sonic, density, and
resistivity) in the context of the Normal Compaction Trend (NCT) theory, and calibrated using direct pressure
measurements (RFT/DST). Additionally, mud weights, pit gains and kicks reported on the daily drilling reports are used to
better constrain the pore pressure profile.
• The orientation and magnitude of SHmax is constrained by observation of drilling induced wellbore failure (breakouts and/or
drilling induced tensile fractures).
• Rock mechanical properties are estimated from log-derived empirical relationships that can be calibrated with laboratory
measurements when available.

Fig. 2 – Example of breakouts and drilling induced tensile fractures observed in a well drilled in the Suria field. From
left to right it is shown: (1) FMI image log, (2) interpreted breakouts (pink) and drilling induced fractures (blue), (3)
caliper, (4) well direction, deviation and pad 1 azimuth, (5) gamma-ray.

1D In-Situ Stress Models: Summary of Results

The 1D stress models were built using logging data from 25 wells and drilling experience from more than 40 wells belonging to
the Apiay, Gaván, Guatiquía, Suria, and Suria-Sur fields. These well-based stress models suggest a present-day strike-slip stress
regime (SHmax > Sv > Shmin). Image logs show the presence of breakouts, but some of them also reveal drilling induced tensile
fractures (Fig. 2) which is an additional constraint feature that demonstrates the strike-slip character of the current stress field
(Zoback, 2007).
The magnitude of the overburden gradient is estimated from vertical integration of density and pseudo-density logs, and
shows a fairly constant value of ~ 18.9 ±0.1 lb/gal at reservoir depths. The minimum horizontal stress was constrained using the
available leak-off tests and mini-fracs. This value ranges between 13 lb/gal and 14 lb/gal at reservoir depths (~10200 ft TVD).
SPE 138869 5

Given the history of uplift that this region has undergone (e.g., Mora et al., 2010), none of the pore pressure methods that deal
with tendency curves can be applied in the area with high degree of confidence. Trend methods are based on changes in porosity.
Uplift increases the porosity of deeply buried formations resulting in false overpressure domains within the uplifted units. In these
cases the most accurate estimate of pore pressure would be that based on the existing drilling experience (e.g., inflows, kicks, and
wellbore failure) and the mud weights used to drill the existing wells. In that matter, a thorough review of the drilling experience
was carried out in more than 40 wells drilled over the last 20 years across the Apiay and Suria oilfields. Based on this information
we have developed 1D pore pressure profiles for each of the wells considered in this study. After reviewing the drilling experience,
the inferred pore pressure profiles showed a similar pattern across these fields, with some differences in the magnitude of the
overpressures. Pore pressure is mostly hydrostatic down to the base of C2 formation, where a gentle pore pressure ramp starts (Fig.
3). A pressure peak tends to occur within T1 and E2 formations, with a reported maximum pressure accros the area of 38 MPa
(11.2 lb/gal EMW at ~9500 ft TVD). The pore pressure profile returns back to a hydrostatic gradient in the reservoir formations.
The virgin hydrostatic pore pressure of the reservoirs has decreased due to production over the 20 last years. Depletion effects on
the horizontal stress magnitudes are modeled when building the 3D geomechanical model.

Fig. 3- Original (non-depleted) stress model in the Apiay field. The relative magnitudes of the three principal stresses
a strike-slip faulting stress regime (Shmin < Sv < SHmax).
6 SPE 138869

The orientation of the maximum horizontal stress was constrained from instances of wellbore failure. Image data collected in
wells across the field showed examples of wellbore breakouts and drilling induced fractures (cf. Fig. 2). Wellbore failure position
around the wellbore suggests an overall E-W orientation of SHmax. This orientation locally varies up to 30 degrees across the field
(Fig. 4).
The magnitude of the maximum horizontal stress is often the most difficult component of the stress tensor to determine. In
this study SHmax magnitude was constrained using the stress polygon diagram of Moos and Zoback (1990). The stress polygon is
constructed based on the assumption that the effective stress ratio cannot exceed that required to cause motion on pre-existing
faults that are optimally oriented with respect to the stress field. Araujo et al. (2010) present a more detailed description of the 1D
modeling work done on the Apiay-Suria wells. The inferred 1D stress models are the main inputs for the 3D model that will be
later used to investigate the stability of the main fault structures.

Apiay
Suria

Fig. 4 – Inferred maximum horizontal stress orientation across the Apiay and Suria areas. The top of Carbonera
Formation is shown as reference.

Three-dimensional modeling
Building a three-dimensional model requires a precise knowledge of the structure, and tectonic evolution of the study area. The
aim of this type of modeling is to integrate the geological and structural knowledge together with the well information to have a
better understanding of the geomechanical properties across the field. Based on the available data, the Apiay, Suria, Suria Sur,
Gaván and Guatiquia fields have been grouped in two modeling areas, that for simplicity will be called Suria and Apiay
geomechanical areas (Fig. 5). The Suria region includes Suria and Suria Sur fields; Apiay includes Apiay, Gaván and Guatiquia
fields.
Structural modeling
Input data for the structural model are the interpreted seismic horizons and major seismically resolvable faults, topography and
well formation markers. Formation markers have been used to build or further constrain the geometry of formation tops where
seismic interpretation was not accurate enough or was not available. The general stratigraphy of the area has been used to define
SPE 138869 7

key horizons in the model.

A further step on the model building process consists of integrating the interpreted fault structures together with the key
horizons. Intersections fault-horizon and fault-fault have been carefully reviewed to assure that structural and geometrical
relationships are held true, and fault throws are properly captured within the structural model. When a water tight model is built,
i.e. there are no gaps between surfaces (faults and/or horizons), the structural model is ready to be used to generate a grid (Fig. 6).
The grid will honor the geological structure and is made of elements with optimal dimensions. That is, the elements of the
grid should be small enough to capture significant variations in the geomechanical attributes, but large enough to speed up
numerical computations. Keeping in mind these two requirements the dimensions of the elements on each modeled formation is as
described on Table 1.

Suria

Apiay

Fig. 5 - Location of the Suria and Apiay 3D blocks.

Model population
Once grid specifications are defined and a structural grid is built, the next level will take wellbore data into the grid and therefore
into the three dimensional area. This is typically done in two consecutive steps: (1) well properties upscaling, and (2) property
propagation.
Wellbore data (e.g., density, acoustic) as well as any log-derived geomechanical attribute (e.g., friction coefficient, UCS)
typically has a high frequency resolution that is far beyond the resolution of any of the grid elements (~ 15 cm sampling rate for a
well log vs. 2-4 m high cell in a structural grid). It is therefore necessary to transfer the log-scale properties to the coarser
resolution grid elements on the vicinities of the well. The value of the property that is going to be placed on each grid element
should be representative enough of the log values within this element. Upscaling properties in a meaningful way usually involves
methods as nearest neighbor, median, mean or most frequent values. To evaluate what is the most adequate method, histograms of
the original and upscaled logs have been conducted looking for reproducibility in both populations. This type of approach
suggested that the arithmetic mean is the upscaling method that better reproduce the original dataset, and thus, it has been used to
generate upscaled logs.
In a second step, upscaled logs are propagated to the three-dimensional space following the structural guidelines imposed by
the grid geometry. There are a number of geostatistical methods that can be applied to generate a volume of properties. It mostly
depends on the type and availability of data as well as on the heterogeneities of the field itself. A thorough review of the wireline
logging data from more than 40 wells across the Apiay and Suria areas demonstrates minor variations on density and acoustic
response of the formations. In general terms, this suggests a fairly homogeneous field, with minor lateral changes in both physical
and rock mechanical properties. Based on this information, a weighted distance method has been selected to model the distribution
8 SPE 138869

of the geomechanical properties within the Apiay and Suria areas.

Amongst the upscaled and propagated properties there are some key geomechanical attributes: SHmax orientation, effective
horizontal stress ratios, UCS, Poisson ratio, friction coefficient, and density.

Fig. 6 – Structural water-tight model.

Three-dimensional stress model

A three-dimensional density volume is the starting point to build any three-dimensional stress model. Integration of density
distribution with depth delivers a volume of vertical stress. One of the advantages of 3D models is that topography is integrated
from the initial stages, and the resulting overburden gradient takes into account any topographic change.
Pore pressure is the second geomechanical attribute to model. The peculiarities of each field usually determine the best
approach to model pore pressure distribution. In this particular case, the 1D geomechanical models suggest a likely hydrostatic
pore pressure gradient that extends down to the C2 unit. Direct pore pressure measurements in sands and drilling experiences
allowed to interpret a pressure ramp starting towards the base of C2, resulting in an overpressurized domain within T1 and E4
formations that gets back to normal pressure at reservoir depths (K1 and K2). These are the guidelines that have been implemented
to obtain an initial 3D pore pressure model for the Apiay and Suria areas. In a further step, direct pore pressure measurements have
been reviewed for all the wells across the study area, to characterize the depletion pattern and generate a likely depleted scenario
for T2, K1 and K2. The actual depletion model assumes the worst case scenario, considering depletion levels for the entire field
equivalent to the lowest pressure measurements available. Depletion is also considered to affect sand rich lithologies, while shales
would remain at virgin pressures. The stress path here is assumed to be 0.6.
Pore pressure and vertical stress are inputs to estimate the magnitude of the minimum and maximum horizontal stresses. The
magnitude of the horizontal stress is determined using the effective stress ratios as establish from wellbore failure modeling. Pore
pressure and overburden relate to the horizontal stresses as follows:

(4)

The volumes of overburden, pore pressure, minimum and maximum horizontal stresses, along with the orientation of the
maximum horizontal stress, characterize the stress tensor across the different formations of the Apiay and Suria areas. This can be
now combined with the seismically resolvable faults in the model to provide assessment on the mechanical stability of these
structures. Examples of the 3D volumes for the vertical and maximum horizontal stresses are shown in Fig. 7.
SPE 138869 9

Fig. 7 – General overview of the Apiay and Suria 3D blocks. The upper view shows the calculated magnitude of vertical
stress (in psi); the view below shows the calculated magnitude of the maximum horizontal stress (in psi).
10 SPE 138869

Fig. 8 – General overview of the Suria and Apiay faults. On the left, critical pore pressure across the analized fault
surfaces. Warm colors represent less stable structures. Cold colors correspond to more stable structures. On the right, pole
diagrams describing fault geometries in the Apiay and Suiria areas and their orientation with respect to the maximum
horizontal stress.

Fault Risk Reactivation Assessment

In order to apply the geomechanical model to assess fault stability, a detailed structural analysis must be carried out to better
understand the 3D fault network within the formations of interest. The inferred effective stress acting on each fracture surface in
the context of the Coulomb Failure Function (CFF) can then be used to determine whether these fractures are active (i.e. shear
stress exceeds normal stress) under the current stress field or under what effective stress they may become active (Barton et al.,
1995). Different fracture orientations will have different tendencies to failure. Similar techniques have been used to asses fault
stability other oilfields around the world (e.g., Finkbeiner et al., 1997; Castillo et al., 2000; Wirprut and Zoback, 2000; Langhi et
al., 2010).
The 3D geomechanical model, and more precisely the stress model, can be used to estimate the stress state on the main faults
of the Suria and Apiay areas, and thus characterize their mechanical integrity. The results of this type of study have a direct impact
on exploration, drilling and production.
Fault geometry
In order to get accurate results from this type of analysis, it is critical to characterize in detail the geometry of faults. In a region
with a homogeneous stress field, as it is the case of the Apiay and Suria areas, the fault orientation with respect to the principal
stress axes, is an important factor that will determine which structures are more prone to be reactivated.
Seismic interpretation of the Apiay and Suria areas has provided valuable information about the faulting pattern. A
preliminary inspection of the fault structures within the studied areas shows a change of structural styles and fracture density
SPE 138869 11

between the areas Apiay and Suria (Fig. 8). The Apiay area consists of highly dipping faults (60-90°), running in a NE-SW
direction and dipping to the NW. To the south, in the Suria area, the main structures rotate towards an E-W orientation as the
dipping becomes more gentle (50-80° N). Assuming that SHmax orientation in the area is roughly E-W, the change in the structural
trend will establish a clear difference on the stability faults between Apiay and Suria.

Fig. 9 – Frequency distribution of critical injection pressures on the fault surfaces within the Apiay and Suria areas.
Tentative risk thresholds are also defined. Shmin denotes the extra amount of pressure required to achive the average
magnitude of the minimum horizontal stress at reservoir depths.

Results
Modeling results are presented in terms of critical injection pressure (in MPa). This is the maximum pressure above the formation
pressure that the faults can handle before slipping in shear. Fig. 8 shows a general view of the results in terms of the more or less
stability of the studied faults (warms colors correspond to less stable structures). Out this general inspection, one can see that there
is a difference between the more stable Apiay area and the less stable Suria area.
Results reveal the presence of certain domains where the critical injection pressure is relatively low (< 10 MPa) and therefore
they may have a high reactivation risk during injection operations. The high risk domains are more frequent within the Suria area;
this is basically imposed by the change in the structural trend. Another important outcome is that there are no negative critical
injection pressures, which means all the faults are stable under the present-day stress field. This observation is consistent with the
fact that some of these faults are seal for the reservoir.
Critical injection pressure histograms have been created for the Suria and Apiay faults to help define different thresholds of
risk (Fig. 9). Faults in the Apiay region show critical injections pressures ranging between 5 MPa and 65 MPa. On the other hand,
Suria faults show a narrower distribution of critical injection pressures: 3 MPa to 40 MPa. As a possible scenario, we have chosen
12 SPE 138869

0 MPa, 10 MPa, and 30 MPa as likely risk thresholds for fault reactivation. This way, a high risk threshold has been defined for
critical injection pressures lower than 10 MPa. Fault domains with injection pressures above 30 MPa are considered as low risk
domains. An intermediate level is established for injection pressures between 10 MPa and 30 MPa. A case of active faulting would
be represented by a 0 MPa injection pressure. Such thresholds can and should be re-evaluated depending on the specifications of
the injection programs.
Fault integrity results for the Apiay area suggest that faults are likely to be more stable, and therefore withstand higher
injection pressures. 75% of the fault domains considered in this analysis requires injection pressures over 38 MPa to induce shear
failure. There are a 10% of these domains that are at a high risk of reactivation, but most of them are located above the reservoir
leves. On the contrary, the Suria area, frequency distribution of critical injection pressures shows that these faults are more prone
to be reactivated. 75% of the population can be reactivated with injection pressures lower than 16 MPa, and 25% is at high risk of
reactivation (<10 MPa). Faults with an associated low risk represent less than 10% of the population.
The magnitude of the average mimimum horizontal stress at the depth of the reservoir is also plotted along with the critical
injection pressure histograms (Fig. 9). It should be noticed, that most faults require lesser pressures than that required to generate a
hydraulic fracture. Therefore, faults can be reactivated before achieving the fracture gradient, which effectively means that the
pressure front generated during any pressure maintenance operation might cause fault slip. An additional corollary is that hydraulic
fracture operations should be planned considering the proximity of these faults, so that when designing the length of the fracture it
should be less than the distance to these faults. If a hydraulic fracture hit a fault, an instantenous communication will be stablished
between the well and the fault, so the fault will see the very same pressure used to propagate the hydraulic fracture. This pressure
in 90% of the cases is higher than the fault reactivation pressures (Fig. 9), so there will be an inmediate slip on the fault, breaching
the seal and promoting fracturing fluids to migrate along the fault as well as reservoir hydrocarbons.
In general, most of the domains at a high reactivation risk are located in the overburden which makes these faults potential
risks for drilling (Fig. 10). It is a common thought that losses take place when the fracture gradient is exceeded due to the
hydraulic fracturing of the formation. However, this is not the only source of losses, and in many cases faults and high bottom-hole
pressures are the reasons for massive mud losses while drilling (e.g. Fernández-Ibáñez et al., 2010). In fault-bounded reservoirs it
is a common practice to drill through some of the main faults to reach the target. If these faults are optimally oriented they could
potentially be reactivated at a lower mud weight than the least principal stress and the upper bound of the operating mud window
could be lowered (i.e. optimally oriented faults could be reactivated before achieving the least principal stress or fracture gradient).

Fig. 10 – Fault domains at a high reactivation risk (< 10 MPa) in the Suria block. Those fault domains located between the
top of T2 and the base of K2 are at a higher risk of reactivation during injection and/or hydraulic fracturing operations.
SPE 138869 13

Conclusions
A three-dimensional geomechanical modeling that takes into account the field geological structure and captures lateral property
variations is here intended to help the design of effective fracturing and water injection programs avoiding risks related to fault
reactivation.
We have used observations of wellbore failure together with logging data and drilling experience to constrain the stress
tensor and rock properties on individual wells distributed across the Gavan, Apiay, Suira, Suria Sur and Guatiquia fields. These
fields were gruped in two major areas, Apiay and Suria, to generate a three-dimensional model. The 1D stress models have been
taken to the three-dimensional space using a structural model as guidance. As a result, we have obtained a static geomechanical
model that honors the structure of the fields and suggests that the Suria and Apiay areas are under a current strike-slip tectonic
regime (SHmax > Sv > Shmin). The orientation of the maximum horizontal stress is E-W and it varies ±30° across the field. An
overpressure has been detected starting at the base of C2 down to the base of E4 formations.
Stress attributes has been projected onto the interpreted fault surfaces to characterize the mechanical integrity of these
structures and evaluate the risks associated to fault reactivation of the reservoir bounding faults during injection and fracturing
opetations. Critical injection pressures have been calculated for each fault, which determines the maximum injection pressures that
the fault can handle. Based on frequency distributions of this property, different risk thresholds has been stablish for the Apiay-
Suria area. None of the studied faults are active under this present-day stress field, which is consistent with the fact that some of
these faults are seal for the reservoir; however, there are some fault segments that are more prone to reactivation. The angle
between the maximum horizontal stress and the geometry of each fault stands as the most contributing factor to the stability of the
fault. Subsequently, the change in structural style between the Apiay and Suria blocks determines a tendency on the last one to be
less stable (higher risk).
The 3D stress model provides constrains on the orientation of hydraulic fractures, so its length can be design to avoid
interaction with any nearby fault. If any hydraulic fracture reaches a fault there is a risk of fault reactivation and hydrocarbon
leakage, but also of fluid migration through the fault to shallower levels. This could also end up in sheared casing of previously
drilled wells through the reactivated fault. It also helps design water injection operations, predicting the maximum pressures that
the faults can withstand before slipping. If water injection for pressure mainteinance is carried out near the high risk faults,
injection pressures should remain below the critical injection pressure threshold proposed in this study.

Acknowledgments
The authors would like to acknowledge and thank both Ecopetrol S.A. and GeoMechanics International – Baker Hughes RDS for
their permission to publish this work.

Nomenclature
CFF - Coulomb Failure Function
EMW - Equivalent Mud Weight
μ - sliding friction coefficient
NCT - Normal Compaction Trend
Pp - Pore pressure
Ppcrit - Critical pore pressure
S1 - Maximun tectonic stress
σ1 - Maximum effective stress
S3 - Minimum tectonic stress
σ3 - Minimum effective stress
σn - Effective normal stress
SHmax - Maximum horizontal stress
Shmin - Minimum horizontal stress
Sn - Normal stress
Sv - Vertical stress
τ - Shear stress
T0 - Tensile strength
14 SPE 138869

References
Anderson, E. M. 1951. The Dynamics of Faulting and Dyke Formation with Applications to Britain, Edinburgh, Oliver and Boyd.
Araujo, E., Alcade, R., Mateus, D., Fernández-Ibáñez, F., Sheridan, J., Ward, C., Brudy, M., Alvarellos, J., Ordoñez, L.Y., and
Cardona, F. 2010. Drilling Optimization Using 3D Geomechanical Modeling in the Llanos Orientales Basin, Colombia,
Paper SPE 138752 presented at SPE LACPEC, Lima, Perú, 1-3 December.
Barton, C.A., Zoback M.D., and Moos, D. 1995. Fluid Flow along Potentially Active Faults in Crystalline Rock. Geology 23 (8):
683-686.
Bayona, G., Cortes, M., Jaramillo, C., Ojeda, G., Aristizabal, J.J., and Reyes-Harker, A. 2008. An integrated analysis of an orogen-
sedimentary basin pair: Latest Cretaceous-Cenozoic evolution of the linked Eastern Cordillera orogen and the Llanos
foreland basin of Colombia. GSA Bulletin 120: 1171-1197, doi: 10.1130/B26187.1.
Breckels, I.M. and van Eekelen, H.A.M. 1982. Relationship between Horizontal Stress and Depth in Sedimentary Basins. SPE J
Petrol Technology 34: 2191-2198.
Byerlee, J.D. 1978. Friction of rocks. Pure & Applied Geophysics 116: 615-626.
Castillo, D.A., Bishop, D.J., Donaldson, I. Kuek, D., de Ruig, M., Trupp, M., and Shuster, M.W. 2000. Trap Integrity in the
Laminaria High - Nancar Trough Region, Timor Sea: Prediction of Fault Seal Failure Using Well-Constrained Stress
Tensors and Fault Surfaces Interpreted from 3D Seismic. APPEA Journal 40 (1): 151-173.
Colmenares, L., and Zoback, M. D. 2003. Stress field and seismotectonics of northern South America. Geology 31 (8): 721-724,
doi: 10.1130/G19409.1.
Fernández-Ibáñez, F., Castillo, D., Firth, A., and Courteney, S. 2010. Assessing Stability of Reservoir Bounding Faults While
Drilling. Paper SPE 132826 presented at SPE ATCE, Florence, Italy, 20-23 September.
Finkbeiner, T., Barton, C.A. and Zoback, M.D. 1997. Relationships among In-Situ Stress, Fractures and Faults, and Fluid Flow:
Monterey Formation, Santa Maria Basin, California. AAPG Bulletin 81, (12): 1975-1999.
Gardner, G.H.F., Gardner, L.W., and Gregory, A.R. 1974. Formation Velocity and Density: The Diagnostic Basics for
Stratigraphic Traps. Geophysics 39: 770-780.
Gomez, A., Jordan, T.E., Allmendinger, R.W. and Cardozo, N. 2005. Development of the Colombian foreland basin system as a
consequence of diachronous exhumation of the northern Andes. Geol. Soc. Am. Bull. 117: 1272-1292.
Holland, M., Brudy,M., van der Zee, W., Perumalla, S., and Finkbeiner, T. 2010. Value of 3D Geomechanical Modeling in Field
Development - A new Approach using Geostatistics, Paper SPEKSA-0059 presented at SPE DGS, Al-Khobar, Saudi
Arabia, 4–7 April.
Jaeger, J.C. and Cook, N.G.W. 1979. Fundamentals of Rock Mechanics. New York: Chapman and Hall.
Langhi, L., Zhang, Y., Gartrell, A., Underschultz, J., and Dewhurst, D. 2010. Evaluating Hydrocarbon Trap Integrity during Fault
Reactivation Using Geomechanical Three-dimensional Modeling: An Example from the Timor Sea, Australia. AAPG
Bulletin 94 (4): 567-591.
Moos, D. and Zoback, M. D. 1990. Utilization of Observations of Well Bore Failure to Constrain the Orientation and Magnitude of
Crustal Stresses: Application to Continental Deep Sea Drilling Project and Ocean Drilling Program Boreholes. Journal of
Geophysical Research 95: 9305-9325.
Mora, A., Parra, M., Strecker, M. R., Sobel, E.R., Zeilinger, G., Jaramillo, C., Ferreira Da Silva, S., and Blanco, M. 2010. The
eastern foothills of the Eastern Cordillera of Colombia: An example of multiple factors controlling structural styles and
active tectonics. Geological Society of America Bulletin 122: B30033.1, doi: 10.1130/B30033.1
Parra, M., Mora, A., Jaramillo, C., Strecker, M.R., Sobel, E.R., Quiroz, L.I., Rueda, M., and Torres, V. 2009. Orogenic wedge
advance in the northern Andes: evidence from the Oligocene-Miocene sedimentary record of the Medina basin, Eastern
Cordillera, Colombia. Geol. Soc. Am. Bull. 121: 780-800.
Sarmiento-Rojas, L.F., Van Wess, J.D. and Cloetingh, S. 2006. Mesozoic transtensional basin history of the Eastern Cordillera,
Colombian Andes: Inferences from tectonic models: Journal of South American Earth Sciences 21: 383-411, doi:
10.1016/j.jsames.2006.07.003.
Schamel, S. 1991. Middle and Upper Magdalena basins, Colombia, in K.T. Biddle, ed., Active margin basins: AAPG Memoir 52:
283-301.
Wiprut, D. and Zoback, M. D. 2000. Fault Reactivation and Fluid Flow along a Previously Dormant Normal Fault in the Northern
North Sea. Geology 28 (7): 595-598.
Zoback, M. D. 2007. Reservoir Geomechanics, New York, Cambridge University Press, 449 pp.