Vacuum Insulated Tubing Computer

Simulation Project

Prepared for:
Mark Godin
PTAC
500 5 Ave SW, Calgary, AB T2P 3L5

Submitted by Shashank Karra

ANSYS Canada Ltd
March 2016

Page | 1
DISCLAIMER: PTAC does not warrant or make any representations or claims as
to the validity, accuracy, currency, timeliness, completeness or otherwise of the
information contained in this report , nor shall it be liable or responsible for any
claim or damage, direct, indirect, special, consequential or otherwise arising out of
the interpretation, use or reliance upon, authorized or unauthorized, of such
information.

The material and information in this report are being made available only under the
conditions set out herein. PTAC reserves rights to the intellectual property
presented in this report, which includes, but is not limited to, our copyrights,
trademarks and corporate logos. No material from this report may be copied,
reproduced, republished, uploaded, posted, transmitted or distributed in any way,
unless otherwise indicated on this report, except for your own personal or internal
company use.
 
Table of Contents
Introduction 1
CFD & Mechanical Transient
Thermal Domain and Mesh Setup 2
Boundary Conditions and Material
Properties 8
Computational Methodology 8
Results and Discussion 10
Conclusions 10
Appendix 11

Page | 2
 Vacuum Insulated Tubing Computer Simulation Project
Introduction
In SAGD, ~100% quality steam is injected through an injector string. Bare tubing is used, resulting in heat
losses from the wellhead to the heel of the well. Therefore, less than 100% quality steam is delivered to
the horizontal section. Vacuum Insulated Tubing (VIT) could replace the bare tubing and significantly
reduce heat losses, resulting in higher quality stream and more heat delivered to the formation.

SAGD operators have experimented with VIT in recent years but adoption is uneven. This is due, in part,
to the difficulty of precisely assessing the energy efficiency and the economic benefits of VIT vs. bare
tubing. Past computer modeling efforts have used basic simulation packages and simplistic geometries.
This simulation in this project used sophisticated CFD and Mechanical Transient Thermal modeling
software by ANSYS.

The objective of this project was to perform a comparative analysis between VIT and bare string tubing
configurations. The simulations provided a comparison of heat losses in both the tubing configurations.

The simulations in Tasks 1 to 4 were steady state fluid flow and heat transfer simulations. These
simulations modeled advection of steam through the tubing, three modes of heat loss namely conduction,
convection and radiation, and condensation of steam into water in the tubing. The simulations in these
four tasks were performed using “ANSYS CFX CFD” software. Task 6 was cancelled to increase the scope
and number of simulations in Task 3.

The simulations in Task 5 and 7 were transient thermal conduction only simulations. These simulations
did not include any fluid flow and modeled just the transfer of heat due to conduction. The simulations in
task 5 and 7 were performed using “ANSYS Mechanical Transient Thermal” software.

Simulation Work Summary

This report summarizes the CFD and transient thermal simulation work spread over six different tasks. A
total of 12 simulations were carried out in all the tasks. The report covers details about the CFD and
mechanical transient modeling methodology, simulated domain, computational mesh and summary of
results and conclusions from the work. The domains created in tasks 1 to 4 were 3D. The domain created
in tasks 5 & 7 were 2D.

1) The four steady state CFD simulations in Tasks 1 and 2 involved studying the geometric sensitivity
of both bare and VIT tubing using steady state CFD simulations. Both bare and VIT tubing were
modeled in concentric and eccentric positions (with respect to the casing).
2) The four steady state CFD simulations in Task 3 involved studying the effect of pseudo-transient
heating of the rock/overburden on heat losses in the tubing. This was carried out in pseudo-
transient manner by changing the distance between the casing and the boundary location of
formation temperature.
3) The two steady state CFD simulations in Task 4 involved studying the effect of varying K value of
VIT tubing on heat losses.
4) The two transient thermal mechanical simulations in Tasks 5 and 7 involved studying the effect
of transient heating of rock/overburden to have a comparison of heat loss vs. time for both bare
and VIT tubing.

Page | 1
The results of each of the above tasks was provided to PTAC and all participating members in the form of
PowerPoint reports.

CFD & Mechanical Transient Thermal Domain and Mesh Setup

A typically CFD and Mechanical transient thermal simulation setup involves creation of a fluid and solid
geometry respectively representing the actual domain dimensions. Meshing the fluid and solid volume,
setting up the simulation, running it to convergence, and finally analyzing the results.

In this project, ANSYS Design Modeler was used to create the simulation domain. ANSYS Meshing was
used to generate a mesh on computational domain, and ANSYS CFX and ANSYS Mechanical transient
thermal were used to run simulations and post process results.

Geometry Creation
The CFD simulation domain created in Tasks 1 and 2 is shown in Figures 1, 2 and 3. The domain for bare
tubing in both concentric and eccentric positions is shown in Figure 1. The domain for VIT tubing in both
concentric and eccentric positions is shown in Figure 2. Rock temperature boundary condition was
applied at 0.01 inch radial distance away from the casing. The dimensions of the domain simulated in
Tasks 1 and 2 are provided in Table 1 in Appendix A.

The contact area between the tubing and casing in eccentric position was measured by zEroCor Tubulars
Inc. in their lab. For bare tubing, contact area was measured to be 7.1% of the circumference of the ID of
the casing and for VIT tubing contact area was measured to be 7.9% of the circumference of the ID of the
casing. These values were used in building the eccentric geometries in all six tasks executed in this project.

Figure 1: Cross Section of Bare Tubing in Concentric Position (left); Cross section of Bare Tubing in Eccentric Position

Page | 2
 Figure 2: Cross Section of VIT Tubing in Concentric Position (left); Cross Section of VIT Tubing in Eccentric Position

Figure 3: 450m Long Simulation Domain containing Tubing, Gas Blanket, Casing and Rock for both Bare and VIT Tubing
Simulations

The simulation domain created for Tasks 3 and 4 is shown in Figures 4 and 5 respectively. Figure 4 shows
domain for bare and VIT in eccentric position with rock temperature applied at 10m radial distance away
from the casing. Similarly Figure 5 shows domain for bare and VIT in eccentric position with rock
temperature applied at 50m radial distance away from the casing. The dimensions of the domain
simulated in these two tasks are provided in Table 1 in the Appendix A.

Page | 3
Figure 4: Simulation Domain of Bare Tubing with Formation Temperature applied at 10m Distance from Casing (left); Simulation
Domain of VIT tubing with Formation Temperature applied at 10m Distance from Casing - Task 3

Figure 5: Simulation Domain of Bare Tubing with Formation Temperature applied at 50m Distance from Casing (left); Simulation
Domain of VIT tubing with Formation Temperature applied at 50m Distance from Casing - Task 4

As mentioned earlier in this report Task 5 and 7 of this study were 2D mechanical transient thermal
simulations. The simulation domain created for Tasks 5 and 7 is shown in Figures 6 and 7. These
simulations were run in ANSYS transient mechanical solver as mentioned in earlier section. A 2D planar
strain domain as created for these simulation. For dimensions of the domain used in these two tasks
please refer to Table 1 in the Appendix A.

Page | 4
Figure 6: Simulation Domain of 2D Mechanical Transient Simulation. The Figure shows Bare Tubing with Formation Temperature
Applied at 30m Distance from Casing

Figure 7: Simulation Domain of 2D Mechanical Transient Simulation. The Figure shows VIT Tubing with Formation Temperature
Applied at 30m Distance from Casing

CFD Mesh Creation

The general CFD mesh created in Tasks 1 and 2 is shown in Figures 8 and 9. A hexahedral mesh was
generated for all the domains namely tubing, gasblanket, casing and rock. Boundary layers were created
on tubing wall to capture flow frictional losses due to non-slipping walls. The tubing wall surface was
assumed to be smooth. Mesh was suitably refined near casing and tubing contact area for eccentric
domain cases.

Page | 5
 Figure 8: The Figure shows Computational Mesh for Concentric Bare Tubing

Figure 9: The Figure shows Computational Mesh for Eccentric VIT Tubing. Similar mesh was generated for Eccentric Bare Tubing
Domain

The CFD mesh created in Tasks 3 and 4 is shown in Figure 10. Similar to the meshes generated in the first
two tasks a hexahedral mesh was generated for all the domains namely tubing, gasblanket, casing and
rock. Boundary layers were created on tubing wall to capture flow frictional losses due to non-slipping
walls. The tubing wall surface was assumed to be smooth. 10 and 50 m rock domain was modeled in Tasks
3 and 4 respectively. Mesh was suitably refined near casing and tubing contact area for eccentric domain
cases.

Page | 6
 Figure 10: The Figure shows Computational Mesh for Eccentric VIT Tubing with 10m of Rock Modeled. Similar meshes were
generated for 50m case as well

The 2D transient structural thermal mechanics mesh created in Tasks 5 and 7 is shown in Figure 11. A
quadrilateral mesh was generated for all the domains namely tubing, gasblanket, casing and rock.

Figure 11: The Figure shows Computational Mesh for 2D Bare Tubing with 30m of Rock modeled. Similar Mesh was generated
for VIT Tubing as well

Page | 7
Boundary Conditions and Material Properties
Tasks 1 and 2
Four CFD simulations were executed in Tasks 1 and 2. These simulations involved modeling of steam
injection through the tubing and heat loss through conduction, convection and radiation. The simulations
also included modeling of condensation of steam into water in the tubing. The boundary conditions used
in Tasks 1 and 2 in this project are given in Table 3 in Appendix B. The material properties used in Tasks 1
and 2 are given in Table 4 in Appendix B.

The key point to note here is that in these simulations 0.01inch thickness of rock domain was modeled.

Tasks 3 and 4
Six CFD simulations were executed in Tasks 3 and 4. These simulations involved modeling of steam
injection through the tubing and three of heat loss through conduction, convection and radiation. The
simulations also included modeling condensation of steam into water in the tubing. The boundary
conditions used in Task 3 and 4 in this project are given in Table 3 in Appendix B. The material properties
used in Tasks 3 and 4 are given in Table 4 in Appendix B

The key point to note here is that in these simulations 10m and 50m thickness of rock domain was
modeled.

Tasks 5, 7
Two mechanical transient thermal simulations were executed in Task 5 and 7. These simulations involved
modeling of transient heat conduction/losses through the tubing into the surrounding rock domain. These
simulation did not include any steam flow modeling. The boundary conditions used in Task 5 and 7 in this
project are given in Table 5 in Appendix B. The material properties used in Tasks 5 and 7 are given Table
6 in Appendix B.

The key point to note here is that in these simulations 30m thickness of rock domain was modeled.

Computational Methodology
Tasks 1, 2, 3 & 4
ANSYS CFX, the CFD solver used for Tasks 1, 2, 3 & 4, uses a finite volume based discretization of the
conservation equations of mass, momentum and energy. The simulation in these tasks involve multiphase
flow of steam and water through the tubing. Steam is modeled as a primary continuous phase in the
domain and water is modeled as a dispersed droplet phase. The conservation equations for mass,
momentum and energy for phase α in multiphase flow are written as:
NP
∂
(rα ρα ) + ∇ • (rα ρα U α ) = ∑ Γαβ
∂t β =1

Γαβ
Where “ ” is the interphase mass flow rate per unit volume between phase α and phase β

Page | 8
 NP
∂
(rα ραU α ) + ∇ • (rα ( ραU α ⊗ U α )) = − rα ∇pα + ∇ • (rα ρα (∇U α + (∇U α )T )) + ∑ (Γαβ
+
U β − Γβα
+
Uα )
∂t β =1

+ +
Γαβ U β − Γβα Uα
Where “ ” represents the momentum transfer induced by interphase mass transfer

As mentioned earlier, the current simulation accounts for condensation of steam into water. To account
for steam condensation in the tubing, the Thermal Phase Change model in ANSYS-CFX was used. The wall
condensation heat and mass transfer rates were determined by a total heat balance in the wall-adjacent
cell assuming thermal equilibrium between the phases. This approach utilised the near-wall vapour
temperature along with local saturation and wall temperatures, which allowed the usage of the thermal
phase change model for wall driven condensation. While this approach does not account for the
nucleation physics happening near the wall or detailed liquid wall film effects, it provides for a good
approximation for the complex process. This approach is best suited for a comparison study, between
different design configurations.

ANSYS-CFX considers the interphase heat transfer due to thermal non-equilibrium across phase
interfaces. The phasic total energy equation is written as
NP
∂ ∂P
(rα ρα hα ,tot ) − rα + ∇ • (rα ραU α hα , tot ) = ∇ • (rα λα ∇Tα ) + rα ∇ • (U α • τ α ) + ∑ (Γαβ
+
hβs ,tot − Γβα
+
hαs ,tot )
∂t ∂t β =1
+ +
Γαβ hβs ,tot − Γβα hαs ,tot
Where “ ” represents the heat transfer induced by interphase mass transfer

For radiation heat transfer, the Monte Carlo model was used in ANSYS-CFX. The Monte Carlo model
assumes that the radiation intensity is proportional to the differential angular flux of the photons and the
radiation field as a photon gas

Tasks 5&7
ANSYS Mechanical, the FEA solver used for Tasks 5 & 7, uses a finite element based discretization of the
conservation equation of energy. ANSYS Mechanical solves the thermal energy equation of the following
form
߲ܶ ߲ܶ ߲ܶ ߲ܶ ߲ ߲ܶ ߲ ߲ܶ ߲ ߲ܶ
ߩܿ ൤ + ‫ݒ‬௫ + ‫ݒ‬௬ + ‫ݒ‬௭ ൨ = ‫ݍ‬ഺ + ൬‫ܭ‬ ൰+ ൬‫ܭ‬ ൰ + ൬‫ܭ‬௭ ൰
߲‫ݐ‬ ߲‫ݐ‬ ߲‫ݐ‬ ߲‫ݐ‬ ߲‫ ݔ‬௫ ߲‫ݔ‬ ߲‫ ݕ‬௬ ߲‫ݕ‬ ߲‫ݖ‬ ߲‫ݖ‬

Where T = temperature (=T(x, y, z, t), ߩ = density, c = specific heat, t = time and ‫ݍ‬ഺ = heat
generation per unit volume

Page | 9
Results and Discussion
The four CFD simulation in Tasks 1 and 2 included bare and VIT tubing in concentric and eccentric
configurations with respect to the casing. The simulations results showed that both in an idealistic
concentric position and in a realistic eccentric position the bare tubing had more heat losses compared to
the VIT tubing. Bare tubing in eccentric position had the highest heat loss. This is because in eccentric
configuration the tubing is in contact with the casing and heat conduction occurs through the contact area
resulting in heat loss and reduction in injected steam quality. On the other hand for VIT tubing heat losses
were minimal in the eccentric position due to the low conductivity material of VIT tubing.

In Task 3 the rock temperature boundary condition was applied at 10m and 50m distance respectively to
investigate the effect of surrounding rock/overburden heating up on heat loss from the injection tubing
in a pseudo transient manner. However the distances of 10m and 50m were too far away and the results
were inconclusive from these simulations. This resulted in modification of simulation methodology which
was executed in Tasks 5 and 7 through transient thermal conduction only simulations.

Task 4 involved studying the effect of VIT K-value on heat loss through injection tubing. It was seen that
decrease in K-value of VIT tubing results in increased heat loss.

The two mechanical transient heat conduction simulations in Tasks 5 and 7 included bare and VIT tubing
in eccentric positions. These transient simulations were run for a total physical run time of five years.
Temperature and heat loss values at different circumferential positions and radial distance (from the
casing) along the rock were recorded with respect to time. The results from these simulation showed that
the heat loss from bare tubing was higher than VIT tubing over the five year period.

Conclusions
All participating producer companies gained valuable insights from the simulations over the performance
of VIT and bare tubing. The simulations showed the importance of considering eccentric placement of
tubing inside the casing and its effect on heat loss. Valuable data from the simulations showing the
temperature and heat loss over five period of time from both bare and VIT tubing was provided to all
participating members of this project.

As next steps all the participating producer companies plan to calculate the economic impact of from heat
loss in the tubing. Future simulation work could involve studying the effect of various parameters such as
tubing diameter, injection steam temperature, rock heterogeneity etc.

Page | 10
Appendix A
Tubing and Casing Domain dimensions simulated in Task 1, 2, 3 and 4

TVD (m) Inclination OD (in) ID (in)

(degrees)

Bare Tubing 450 10 3.5 2.99

VIT Tubing 450 10 4.5 2.99

Casing 450 10 9.625 8.9

Table 1: Domain Dimensions for Casing and Tubing used in CFD simulations in Tasks 1 to 4

Task 5 and 7 were 2D mechanical transient thermal simulations. The dimensions for casing and tubing
used in these simulations are given below

OD (in) ID (in)

Bare Tubing 3.5 2.99

VIT Tubing 4.5 2.99

Casing 9.625 8.9

Table 2: Domain Dimensions for Casing and Tubing used in Transient Thermal Simulations in Tasks 5 and 7

Page | 11
 Appendix B
Boundary Conditions used for CFD simulation in Tasks 1, 2, 3 and 4 are given below

Steam Injection Flow 250

rate (m3/day)

Steam Injection 3
pressure (Mpa)

Steam Injection 434

Temperature (C)

Steam Quality 0.957

Rock Temperature (C) 10

Table 3: Boundary Conditions used in CFD simulations in Tasks 1 to 4

Material properties used in CFD simulations in Task 1, 2 and 3 are given below

Bare Tubing VIT Tubing Casing (Steel) Gas Blanket Athabasca Oil
(Steel) (Nitrogen) Sands *

Density (kg/m3) 7850 3000 7850 808.4 2073

Thermal Conductivity 44 0.01 44 0.05 2.5

(W/mC)
Table 4: Material Properties used in CFD simulations in Tasks 1 to 4

For task 4 (which involved CFD simulations comparing different K-values of VIT tubing) K-values of 0.006
and 0.1 were used.

Page | 12
 Boundary Conditions used in mechanical thermal transient simulation in Tasks 5 and 7 are given below

Steam Injection 250

Temperature (C)

Rock Temperature (C) 10

Table 5: Boundary Conditions used in Transient Thermal Simulations in Tasks 5 and 7

Bare Tubing VIT Tubing Casing (Steel) Gas Blanket Athabasca Oil
(Steel) (Nitrogen) Sands *

Density (kg/m3) 7850 3000 7850 808.4 2073

Thermal Conductivity 44 0.01 44 0.05 2.009

(W/mC)

Specific Heat Capacity 434 750 434 1040 900

(J/kgk)
Table 6: Material Properties used in Transient Thermal Simulation in Tasks 5 and 7

Page | 13