Analyzing Geomechanical Effects while Drilling Salt

Wells through Numerical Modeling

Freddy Mackay1, Fabrcio Vieira Cunha Botelho2, Nelson Inoue3 and Sergio
Augusto Barreto da Fontoura4
1,3,4

PUC-Rio University/Group of Technology and Petroleum Engineering 2 PETROBRAS Petrleo Brasileiro S. A.

Abstract: Santos Basin is one of the most promising basins of Brazil, recently it was discovered
light crude oil of 30 API (American Petroleum Institute), this reservoir of high productivity is
located below a salt layer of two thousand meters of thickness. Salt also known as an evaporite
rock is found in many hydrocarbon basins around the world. Evaporites are sediments formed
initially from minerals dissolved in water, the most common are: halite, gypsum, and anhydrite.
These minerals are found in areas that passed by a geological time of high evaporation or
precipitation. Evaporites in general have the structure of a dome, formed when a thick layer of
salt found in the bottom begins to crossover the superior layers vertically, its process delays
millions of years. The presence of saline structures takes into favorable conditions for creating a
trap for hydrocarbons, increasing the success of probability in oil and gas exploration. In salt
drilling the main problem is the closing of the well or known as squeeze or salt pinch, this
phenomenon provokes the imprisonment of the drillstring also known as stuck pipe. Evaporite
rock behavior is defined through a creep model. The creep model is the term used in the
bibliography to define the tendency that a material has to move or to deform permanently to
relieve stresses. Strains take place due to extended levels of stress; this stress is less than the
rupture stress. In this matter it was chosen a numeric tool to simulate the behavior of salt drilling
and the effects that this will have due to the overburden stress and drilling fluid pressure. This tool
is Abaqus a program of finite elements in 3D (three dimensions). The objective of this work is to
determine the effect of the fluency (creep) in salt during and after the drilling of the well, through
numeric simulation. At first it is shown a validation of Abaqus finite element program comparing
results from literature and then it is shown a simulation of a 3D vertical salt wellbore.
Keywords: Creep, Drilling, Geomechanics, Evaporites and Wellbore.

1. Introduction
Among offshore basins of Brazil, Campos and Santos basins located Southeast of Brazil,
nowadays, are receiving considerable attention from the petroleum industry due to the great results
of exploration and production of hydrocarbons. Campos basin is responsible for more than 80% of
Brazils oil production and Santos basin has enormous growth perspectives due to the discovery of
new light crude oil. Not long ago, Petrobras confirmed the discovery of light crude oil of 30 API
in Santos Basin; finding a reservoir of high productivity, located below a salt layer of two
2008 Abaqus Users Conference

thousand meters of thickness ("pre-salt") 2,000 m (6561,7 ft). The presence of saline structures in
these basins forms favorable conditions to seal petroleum fluids, increasing the success and
probability in oil and gas exploration. In this context, evaporite rock is a caprock because it
doesnt allow fluids to flow through it, presenting very low permeability. On the other hand, many
operational problems in drilling salt structures have been brought up to light, some of the most
common problems registered by the petroleum industry are stuck pipe due to salt pinch and
wellbore collapse due to inadequate drilling fluid design. Besides these important remarks, the
strains of sediments adjacent to salt structures, combined with the stresses of disturbances caused
by the presence of salt usually brings a risk in drilling, forming a transition zone where the
principal characteristic is the instability of the well and problems of lost circulation. These
problems create great challenges and generate the opportunity of evolution in the oil industry. In
this way we can say that the complexity of saline bodies and the deep reservoirs request not only
high development costs, but also an innovative technology to reach the production fields, being
necessary the use of special procedures for drilling through evaporites. An example of this
particularity was studied by Willson & Fredrich (2005), Fredrich et al. (2003), Poiate Jr.et al.
(2006) and others. These authors presented and showed how these uncertainties close and through
salt can be included in the geomechanical project of the well and in the weight of the necessary
drilling fluid. The objective of this work is to determine the effect of creep in salt during and after
the drilling of the well, this is done by analyzing the behavior of salt creep in a salt well with and
without the interaction of weight of the drilling fluid. This evaluation is made through numeric
simulation by a finite element commercial program.
1.1

Evaporites in Brazil

Evaporites are sedimentary rocks that present saline minerals layers, being the principal
component halite. These are deposited directly of brines in conditions of strong evaporation and
precipitation of sedimentary, restricted, hot and subsidence basins. Those deposits could be of
continental or marine origin; where there is a periodic contribution. The formation of evaporites
are due to trapped lagoons in tropical climates with strong and continuous evaporation, along with
systematic or intermittent affluence of salt seawater and with little or any contribution of clastic
sediments. The precipitation of salt happens when the solute reaches the point of saline saturation
of the respective component. The deposition of saline layers take place as a sequence or
succession of progressive salinization of the deposition of the basin. The order of deposition is
from less soluble salts to the most soluble salts; for instance, gypsum and anhydrite in lower
layers, and halite, silvite and carnalite in higher layers. Evaporites are found in several
hydrocarbon basins around of the world. Oliveira et al. (1985) presented a geological study,
including aspects of evaporites mainly in Campos Basin. According to this work, the formation of
evaporitos in Brazil took place about 135 million years ago, in other words, in the Lower
Cretaceous. The process of continental separation created gulfs, previous to total separation (open
sea), allover the coast. This propitiated favorable conditions of restriction of seawater flow. All of
this process associated to environmental conditions such as dry and hot weather, airing,
evaporation, feedings of water source and morphologic restriction were favorable to the formation
of evaporite deposits in the Brazilian coast. According to a proposed model, the vertical ascension
of salt bodies called halokinesis, has its origin in local salt deposits in each of the corresponding
lagoons along the basin. This phenomenon is the salt penetration through upper layers of dense
rock producing domic structures. As a result of continental separation, the condition of the
2

2007 ABAQUS Users Conference

environment stopped being restricted, provoking the deposits of sediments in open sea. Santos and
Campos basins are complex examples of saline structures. The seismic stratigraphic interpretation
and tectonics in these areas are well understood, this is a result of decades of exploration. These
sedimentary basins were studied by Chang et al. (1992), they were formed in the Mesozoic era.
Marlim field in Campos Basin, for instance, is a field of gigantic oil reserves where originally had
a volume of approximately 10 billion bbls of oil, becoming the largest offshore oilfield known in
the world. The study of diapiric structures in saline rocks is very important because evaporites are
important seals for petroleum fluids, they are excellent caprocks. On the other hand, structures in
proximity to salt diapirs are complex. This is because the penetration of salt bodies (halokinesis)
causes the deformation of upper layers of dense rocks. The consequence of this is the formation of
intense shear around the salt dome, fractured areas and failure zones on the top of the salt dome.
These are some of the factors that complicate the drilling of salt surrounding areas.

2. Numerical modeling
2.1

Creep Model of the Finite Element commercial program

The creep model presented and used in the finite element commercial program follows the time
hardening form of the Power-law creep model. The time hardening form according to the
commercial program has the following equation:

& cr = A q n t m

Equation 1

Where is the uniaxial equivalent creep strain rate, q is the uniaxial equivalent deviatoric stress
(Mises equivalent stress), t is the total time, and A, n and m are parameters defined by the user as
functions of temperature. In order to simulate creep on the commercial program, elastic
parameters must be introduced E (Youngs modulus) and (poissons ratio). This is because it is
considered that an element that has a creep behavior must also have an elastic behavior. It is not in
the interest of the this paper to discuss the several creep models found in the literature. Although it
is important to remark that the time hardening form of the Power law creep model suits well for
these simulations.
2.2

Validation of the commercial finite element program

In order to validate further results, it is necessary to match known literature results in the the finite
element program. The first validation was made through Bradleys formulation for a circular
wellbore, analyzing the elastic response of the well. The second validation was the 3D response of
the program by obtaining similar results of Fredrich et al. (2003) experiencies in determining 3D
in-situ stress in salt bodies. For the first validation, the case of study is the drilling of an oil well of
0,216 m (8") of diameter in a section in two dimensions (2D), located in the salt layer. This
section has the particularity to represent a quarter of the wellbore. The quarter of the wellbore has
the purpose of reducing the mesh and computer costs, see Figure 1. The depth of study is 2700 m
(8858,3 ft) below the sea level.
2008 Abaqus Users Conference

Figure 1. Finite Element 2D mesh.

The equations 2 and 3 formulated by Bradley (1979) for a circular wellbore are:

3a4 4a2
x + y a2 x y 3a4 4a2
a2
1 2 +
1+ 4 2 cos(2 ) + xy1+ 4 2 sin(2 ) + pw 2
r
r
r
2 r 2 r
r

r =

3a4
x + y a2 x y 3a4
a2
1 + 2
1+ 4 cos(2) xy1+ 4 sin(2) pw 2
r
2 r 2 r
r

Equation 2

Equation 3

Where r is the normal radial stress, x is the in-situ stress in the x axis, y is the in-situ stress in
the y axis, a is the radius of the wellbore, r is the distance from the the center of the wellbore to the
specified element, = is the angle measured counterclockwise from x axis of plane x-y, is the
tangential stress and pw is the pressure due to the drilling fluid.
The principal characteristics of the case of study are presented in Table 1:
Table 1. Principal characteristics of the 2D validation case.

Material
Sea water

Density
3
3
1027 kg/m (64,11 lb/ft )

Other
sediments
Salt layer

1 psi/ft = 2306,66 kg/m

3
(144 lb/ft )
3
2160 kg/m
3
(134,84 lb/ft )

Depth
0 m to -500 m
(0 ft to -1640,4 ft)
-500 m to -2500 m
(-1640,4 ft to -8202,1 ft )
-2500 m to -3000 m
(-8202,1 ft to -9842,52 ft)

2007 ABAQUS Users Conference

In order to transfer the total pressure to the studied section located at 2700 m (8858,27 ft) below
the sea level (-2700 m), it is necessary to separate the stresses generated by the three areas (sea
water, other sediments and the salt layer itself). The first area called sea water represents the
pressure caused by the weight of the ocean. The area named other sediments stands for the rock
layers situated above the salt, these layers have a density of 2306,66 kg/m3 (1 psi/ft), this value is
accepted as an ideal gradient to simulate the overload of rock formation in the petroleum industry.
Finally the salt layer, which represents the salt rock. In this case it was chosen that Halite would
represent best the salt layer. Halite is a component of the salt rock and its density is 2160 kg/m3
(134,84 lb/ft3), Medeiros (1999). These formulations were used for the plane strain case. In this
simulation the program used only the elastic step. The finite elements analyzed are those found in
the lower part of Figure 1 along axis 1. In order to study the behavior of the salt well this study
was analyzed under plane strain case, the mesh was built in 2D. The stresses and the
displacements in the wellbore were analyzed considering the use of the drilling fluid and not
considering it. The salt layer was modeled considering the creep behavior before explained. The
general explanation of the numerical simulation procedure is as follows. To simulate the creep
behavior of the salt well, a mesh of finite elements was built in 2D. This mesh has the main
characteristic of one quarter of the total dimensions of the wellbore, the size of the mesh is of 20
times the radius, as the radius is 0,108 m (4 "), 20*r, this would be 2,16 m (7,09 ft), the
dimensions of the mesh are of 2,16 m * 2,16 m. It has 234 plane strain elements (CPE4) and 248
nodes. The characteristic of plane strain elements is that they assume that the out of plane strain is
zero. These elements were used in this simulation because they match the 2D criteria of a section
wellbore in assuming that the out of plane strains are zero. The reason for which it was adopted 20
times the radius is that in this distance the stresses are equal to in-situ stresses, this fact was
verified through previous simulations in the commercial finite element program, and after taking
this consideration this mesh size was adopted. In figure 1 the 2D finite element mesh is shown for
the simulation of the drilling of the well and in Figure 2 it is shown a zoom of Figure 1, this with
the objective of showing the type of elements that were used in the simulation.

Figure 2. Zoom of the 2D mesh.

2008 Abaqus Users Conference

From Figures 1 and 2, it can be seen that there are two types of finite elements used to simulate the
salt layer, explicitly the wellbore and the rest of the salt rock. In the circle area, in other words, the
quarter of the well from the center of the well to the perimeter of the well (see Figure 2) triangular
finite elements of 3 nodes were used and in the rest of the mesh quadrilateral finite elements of 4
nodes were used. Once again the analysis as quoted before is set by the plane strain method. The
axes represented by the numbers 1, 2 and 3 correspond to x, y and z axes. The nodes located at the
vertical left end of figure 1 and the nodes located at the bottom of figure 1 are set to no
displacements at all, simulating that it is the continuity of wellbore. After simulating the response
of the wellbore due to drilling, it has been found that the results proposed by Bradley (1979)
coincide with the results of the commercial program (see Figure 3). In Figure 3 there is a graphic,
showing the response for radial stresses and tangentials stresses along the elements, these have
almost the same values. For the second validation it was chosen the work done by Frederich et al.
(2003) obtaining similar results in determining 3D in-situ stress in salt bodies.

Figure 3. Elastic analysis of Bradleys formulations vs Abaqus Finite Element

Program.
Several body shapes were simulated and the same results were found. One example of this is the
simulation of a 2000 m (6561,7 ft) salt sphere at 4000 m (13123,3 ft) of depth (top of the sphere).
This example was using a mesh of 4625 nodes and 24119 4-node continuum elements. The stress
distribution of the numerical simulation of the salt sphere for the von Mises stress, the vertical and
horizontal stress agrees with the results presented by Frederich et al. (2003), see Figure 4, 5 and 6.

2007 ABAQUS Users Conference

Figure 4. Validation of the 3D capability of Abaqus. Von Misses stresses of a salt

sphere. Similar results (Frederich, 2003).

Figure 5. Validation of the 3D capability of Abaqus. Vertical stress of a salt sphere.

Similar results (Frederich, 2003).

2008 Abaqus Users Conference

Figure 6. Validation of the 3D capability of Abaqus. Horizontal stress of a salt

sphere. Similar results (Frederich, 2003).

3. Cases of study
3.1

2D salt well, creep simulation

After validating the finite element program, new experiences using the program can be
accomplished. Using the same input values for the 2D validation case, plane strain case, the
following results were obtained; see figure 7 and figure 8.

Figure 7. 2D salt well creep stress behavior.

8

2007 ABAQUS Users Conference

Figure 8. 2D salt well creep behavior for different drilling fluid weights.
Figure 7 presents the analysis of the 2D salt well stress behavior studying the effect of creep. This
simulation uses the creep step of the program. This means that three steps are used in this
simulation: step 1, introduces the in-situ stresses, step 2 is the elastic response after drilling, and
step 3, shows the creep behavior of the salt well. The elements analyzed in Figure 7 are those
located at the bottom of Figure 1 along axis 1. These elements have the characteristic that the
integration points are located near the center of each element, forming 2,5 with axis 1. The values
of stresses were obtained through each of these integration points. Also in figure 7, it is shown the
elastic response and the creep response of the salt well after the drilling of the wellbore. It is
important to notice that in this simulation, the pressure of drilling fluid wasnt taking into account,
this is because it was important to understand how the stresses vary through time in a wellbore
with no drilling fluid. Other results taking account of the drilling fluid were analyzed also, but are
not shown here, this for the sake of showing the most interesting results, although this is not the
real life case. The elastic response of the salt wellbore is the response of the wellbore immediately
after the well has been drilled, this response is shown through several values of stress of the
elements located along axis 1 (elements located in the first row of the bottom of Figure 1). In
figure 7 it is shown curves of the radial and tangential stress of these elements. After the elastic
response, follows the creep response shown as Elastic + Creep response this step shows how
stresses have changed after time passed, arrows show the trajectory of stresses through time for
each element. Only for the purpose of an academic study the final creep time was set to 10E11
seconds. Analyzing the tangential stress curves only, it is possible to see that up to a distance of
approximately 5 times the radius 5*R (distance from the center of the wellbore) the values of the
tangential stress stop decreasing, and from there on they increase in there values through time.
Another interesting fact is that the radial stresses decrease their values through time in all the
studied elements, until 20*R which is the distance where stresses stabilizes, e.g. are equal to field
stresses. In figure 8 it is shown the graphic of the wellbore closure through time for different
values of the drilling fluid pressure, these values vary from 9 ppg (0,47 psi/ft) to 13 ppg (0,67
2008 Abaqus Users Conference

psi/ft). In the numerical simulation this fluid pressure is taking into account as non-penetrating
only analyzing the geomechanical response of the wellbore. In this figure the node of study is
node No. 119 (see Figure 2). This node is located at the edge of the wellbore. It is seen from figure
8 that different weights of drilling fluid have different response on the wellbore closure. At the
end, a lower drilling fluid weight provokes a higher wellbore closure.
3.2

3D vertical salt well drilling simulation

In order to simulate the drilling of a 3D vertical salt wellbore a mesh of 2460 nodes and 1880 8node hexahedral elements with reduced integration (C3D8R) was created. This mesh can be seen
in Figure 9. Note that the elements in the center are the elements that will simulate the wellbore.

Figure 9. Finite Element mesh for the 3D vertical salt wellbore drilling simulation.
The purpose of this mesh is to represent the drilling of 14 m of a salt wellbore located at 4000 m
(13123,3 ft) deep. The simulation of the drilling was divided in 4 stages of 3,5 m (11,5 ft) (see
Figure 10). In figure 10, 0 m represents the top of the mesh this point has 4000 m (13123,3 ft) of
depth, 3,5 m (11,5 ft) represents 4003,5 m (13131,48 ft) and so on. Node No. 18 has 4003,5 m
(13131,48 ft) of depth and node No. 9 has 4010,5 m (13154,44 ft) of depth. Nodes No. 18 and 9
will be studied further on.

10

2007 ABAQUS Users Conference

Figure 10. Description of the 3D vertical salt well.

The overburden weight passed at the top of the mesh, represents the weight of the rock mass, this
overburden was calculated using the 1 psi/ft gradient. The total stress at the top of the mesh would
be 90520000 Pa (13128,82 psi) similar procedure was adopted to pass the stresses to the
sideburden of the mesh (in Figure 9, these sides are the ones located at the back), this also taking
into account that the horizontal to vertical ratio of stress is 1. The drilling fluid weight is 11 ppg
(0,57 psi/ft), that produces a pressure of 51720000 Pa (7501,35 psi = 51,7 MPa). Three steps were
used to simulate the drilling, 1 the in-situ stresses, 2 the elastic response of the wellbore after
being drilled and the immediate pressure of the drilling fluid to the faces of the wellbore and
finally 3 the creep response of the salt mass. Steps number 2 and 3 were used to simulate the rest
of the drilling stages. Figures 11 and 12 show the results of the wellbore closure analyzing the
presence of drilling fluid and not. Both figures 11 and 12 study the reponse of nodes 18 and 9.
Figure 11 shows the response of wellbore closure vs the depth of penetration. Figure 12 represents
the results of wellbore closure through time (in hours). From these two figures analyzing only the
behavior of node No. 18 it seen that the elastic response begins when the drill bit reaches this
depth this elastic response is immediately, after the creep behavior starts to come in evidence.

2008 Abaqus Users Conference

11

Figure 11. 3D salt well, elastic + creep behavior for nodes No. 18 and 9. Notice the
elastic behavior first and then the creep behavior.

Figure 12. 3D vertical salt well, elastic + creep behavior for nodes No. 18 and 9.
Analyzing wellbore closure vs time.
Figure 13 and 14 shows the stress response of the vertical 3D drilling simulation of a salt well
considering the creep behavior. In Figure 13 it is shown a red line, this line marks the elements
which are analyzed. Figure 14 includes the radial and tangential stress response; this figure is
similar to Figure 7. Comparing Figure 7 and figure 14 it is seen that radial and tangential
demonstrate the same behavior in their trajectories through time.

12

2007 ABAQUS Users Conference

Figure 13. Finite Element mesh for the 3D vertical salt wellbore drilling simulation.
Elements marked in red are analyzed in Figure 14.

Figure 14. 3D salt well creep behavior response of radial and tangential stresses.

2008 Abaqus Users Conference

13

4. Conclusions
The time hardening form is the simplest form of the power law model and it has been
demonstrated that it suits well to develop the results of the salt wellbore drilling simulation.
The A, n and m parameters of the time hardening power law presented will vary depending on the
depth (earth pressure - overburden) and temperature (temperature gradient). In this paper the
parameters were calibrated from Costa et al. (2005).
Through the validations it was possible to acknowledge that the finite element commercial
program can reproduce accurate results.
In the numerical simulations presented in this paper, 2D and 3D results were obtained regarding
mainly the wellbore closure and the induced stresses caused after the drilling. Then it was possible
to state that the cause of a rapid wellbore closure is a low weight of drilling fluid. This is an
essential data to know what will be the specific time lapse of the casing, before the well suffers a
complete salt pinch, it is also important to prevent stuck pipe. Another interesting response is the
trajectories of stresses trough time. These trajectories at some point estabilize, this depends if they
are tangential or radial stresses. Tangential stresses decrease their values through time in up to five
times the radius and then they increase their values. Radial stresses decrease their values through
time in all the radial extension up to almost 16 times the radius. At a length of 20 times the radius
(or before), starting from the center of the wellbore, they behave like the continuum medium
(equal to the far field stress).
These results present an initial stage of the research of salt-evaporites behavior. In further research
works the objective is to determine how parameters such as depth, temperature, evaporite type and
others influence results.

5. References
1. Barker, J.W. & Meeks, Estimating Fracture Gradient in Gulf of Mexico Deepwater, Shallow,
Massive Salt Sections. SPE 84552, October 2003.
2. Bradley, W.B., Failure of Inclined Boreholes. J. of Energy Resources Tech., 1979.
3. Chang, H.K. Kowsmann, R. O.; Figueiredo, A. M. F.; Bender, A. A., Tectonics and
stratigraphy of the East Brazil Rift system: an overview Tectonophysics 213, p 97-138,
October 1992.
4. Costa, A. M.; Poiate Jr., E.; Falco, J. L. & Coelho L. F. M., Triaxial Creep Tests in Salt
Applied in Drilling Through Thick Salt Layers in Campos Basin Brazil IADC/SPE 92629.
February 2005.
5. Dowling, N. E., Mechanical Behavior of Materials. Engineering Methods for Deformation,
Fracture, and Fatigue Second Editon, 1999.
6. Dusseault, M. B., Alaska Rocks Course on Earth Stresses and Drilling Rock Mechanics Module H - Stresses and Drilling in and Around Salt Structures University of Waterloo and
Geomec a.s., 2005.
14

2007 ABAQUS Users Conference

7. Frederich, J.T.; Coblentz, D.; Fossum, A.F. & Thome, B.J., Stress Perturbations Adjacent to
Salt Bodies in the Deepwater Gulf of Mexico. SPE 84554. October 2003.
8. Kupfer, D.H., Boundary shear zones in salt rocks IV Symp. Salt, Houston, USA.1974,
p215-225.
9. Medeiros, F. A. S., Anlise do Comportamento de Colunas de Revestimento Frente
Movimentao do Sal em Poos de Petrleo Dissertao de Mestrado. PUC-Rio, 1999.
10. Munson, D. E.; Devries, K. L., Development and Validation of a Predictive Technology for
Creep Closure of Underground Rooms in Salt Seventh International Congress on Rock
Mechanics. Vol 1, pp 127-134, Aachen/Deutschland, 1991.
11. Oliveira, J. E.; Idagawa, L. S.; Nogueira, E. C. Evaporitos na Bacia de Campos, Aspectos
Geolgicos e Problemas de Perfurao PETROBRAS/CENPES-475, 1985.
12. Poiate Jr., E.; Costa A. M & Falco J. L., Well Design for Drilling Through Thick Evaporite
Layers in Santos Basin Brazil IADC/SPE 99161. February 2006.
13. Willson, S. M. & Fredrich J. T., Geomechanics Considerations for Through an Near Salt
Well Design, SPE 95621, October 2005.

6. Acknowledgment
The authors would like to express their gratitude to the Program of Human Resources of ANP of
Brazil (PRH-ANP Agncia Nacional do Petrleo, Gs Natural e Biocombustveis) and Pontifical
Catholic University of Rio de Janeiro (PUC-Rio), for the scholarship granted to Mr. Mackay
during the course of this work.

2008 Abaqus Users Conference

15