University of Calgary PRISM: University of Calgary's Digital Repository Graduate Studies The Vault: Electronic Theses and Dissertations 2016-01-08 Performance of Steam Assisted Gravity Drainage in Thin Oil Sand Reservoirs: Well Pair Configuration Zohrehvand, Shiva Zohrehvand, S. (2016). Performance of Steam Assisted Gravity Drainage in Thin Oil Sand Reservoirs: Well Pair Configuration (Unpublished master's thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/27300 http://hdl.handle.net/11023/2737 master thesis University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca
157
Embed
Performance of Steam Assisted Gravity Drainage in Thin Oil Sand … · 2019. 4. 24. · UNIVERSITY OF CALGARY . Performance of Steam Assisted Gravity Drainage in Thin Oil Sand Reservoirs:
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
University of Calgary
PRISM: University of Calgary's Digital Repository
Graduate Studies The Vault: Electronic Theses and Dissertations
2016-01-08
Performance of Steam Assisted Gravity Drainage in
Thin Oil Sand Reservoirs: Well Pair Configuration
Zohrehvand, Shiva
Zohrehvand, S. (2016). Performance of Steam Assisted Gravity Drainage in Thin Oil Sand
Reservoirs: Well Pair Configuration (Unpublished master's thesis). University of Calgary, Calgary,
AB. doi:10.11575/PRISM/27300
http://hdl.handle.net/11023/2737
master thesis
University of Calgary graduate students retain copyright ownership and moral rights for their
thesis. You may use this material in any way that is permitted by the Copyright Act or through
licensing that has been assigned to the document. For uses that are not allowable under
copyright legislation or licensing, you are required to seek permission.
Downloaded from PRISM: https://prism.ucalgary.ca
UNIVERSITY OF CALGARY
Performance of Steam Assisted Gravity Drainage in Thin Oil Sand Reservoirs: Well Pair Configuration
by
Shiva Zohrehvand
A THESIS
SUBMITTED TO THE FACULTY OF GRADUATED STUDIES
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
GRADUATE PROGRAM IN CHEMICAL AND PETROLEUM ENGINEERING
Success of Steam Assisted Gravity Drainage (SAGD) depends on reservoir properties and
operational parameters. Here, both areas are targeted and performance of SAGD in thin
oil sand reservoirs with changing the well configuration is studied. Specifically, the
influence of the injection and production wellpair configuration as well as the number of
injector wells in a homogeneous formation with thicknesses of 5, 7, and 10m were
investigated. The wellpairs were relocated to make different patterns where the
spacing between injection and production wells was changed. SAGD performance was
assessed numerically and the cumulative steam oil ratio, oil production, heat loss, and
oil recovery factor were compared. The results suggest that the horizontal and vertical
distances between injectors and the producer well, their locations from over or
underburden and their alignments affect the performance of SAGD operation. The
results also show that addition of an offset injector well can be beneficial.
iii
Acknowledgements
First and for most I am very grateful to my supervisor Dr. Ian Gates. Thank you Ian for
being a great mentor and an incredible human being.
Thank you Dr. Bahareh Khansari for your remarkable comments and great friendship.
Thank you Jacky Wang for the invaluable discussion.
I highly appreciate the financial support of “Werner Graupe” scholarship and Computer
Modeling Group Ltd. (CMG) for providing the reservoir simulator CMG STARSTM.
iv
To my family
Thank you for your unconditional love and support
v
Table of Contents
Abstract...............................................................................................................................ii Acknowledgments..............................................................................................................iii Dedication..........................................................................................................................iv Table of Contents................................................................................................................v List of Tables.....................................................................................................................viii List of Figures.....................................................................................................................ix List of Symbols, Abbreviations and Nomenclature............................................................xi CHAPTER 1: INTRODUCTION...............................................................................................1
1.1. Statement of the Problem........................................................................................2 1.2. Objectives of the Thesis............................................................................................3 1.3. Research Methodology.............................................................................................3 1.4. Outlines of the Thesis...............................................................................................4
CHAPTER 2: LITERATURE REVIEW.......................................................................................6 2.1. Oil Sands Recourses..................................................................................................6 2.2. Chemistry of Heavy Oil and Bitumen........................................................................9 2.3. EOR Methodologies................................................................................................11
2.5. Thin Oil Sands Reservoirs........................................................................................24 2.6. Well Spacing and Configuration..............................................................................26 2.7. What is Missing in the Literature? .........................................................................31 CHAPTER 3: PERFORMANCE OF STEAM ASSISTED GRAVITY DRAINAGE IN THIN OIL
3.3.1. Model H10..................................................................................................37 3.3.2. Model H7....................................................................................................40
3.3.3. Model H5....................................................................................................42 3.4. Results and Discussion............................................................................................45
3.4.1. Model H10..................................................................................................45 3.4.1.1. Cumulative Steam-to-Oil Ratio....................................................45 3.4.1.2. Cumulative Produced Oil.............................................................49
3.4.2. Model H7....................................................................................................57 3.4.2.1. Cumulative Steam-to-Oil Ratio....................................................57 3.4.2.2. Cumulative Produced Oil.............................................................60 3.4.2.3. Oil Recovery Factor......................................................................62 3.4.2.4. Cumulative Heat Loss..................................................................64
3.4.3. Model H5....................................................................................................66 3.4.3.1. Cumulative Steam-to-Oil Ratio....................................................66 3.4.3.2. Cumulative Produced Oil.............................................................69 3.4.3.3. Oil Recovery Factor......................................................................71 3.4.3.4. Cumulative Heat Loss..................................................................73
3.4.4. Best Case Scenarios....................................................................................76 3.4.4.1. Cumulative Steam-to-Oil Ratio....................................................76
3.4.4.2. Cumulative Produced Oil.............................................................77 3.4.4.3. Oil Recovery Factor.....................................................................78 3.4.4.4. Cumulative Heat Loss..................................................................79 3.4.4.5. Temperature Distributions and Well Pairs Arrangement............81
3.5. Conclusions.............................................................................................................83 CHAPTER 4: PERFORMANCE OF STEAM ASSISTED GRAVITY DRAINAGE IN THIN OIL SANDRESERVOIRS: Well-pair Configuration in a Single Producer-Double
4.3.1. Model H10-2Inj...........................................................................................91 4.3.2. Model H7-2Inj.............................................................................................94 4.3.3. Model H5-2Inj.............................................................................................96
4.4. Results and Discussion............................................................................................98 4.4.1. Model H10-2Inj...........................................................................................99
4.4.4. Best Cases for the Single Producer-Dual Injector Models.........................120 4.4.5. Best Cases for the Single Injector-Single Producer and
Table 2-1 API gravity classification of petroleum oil (API, 2013)................................7 Table 2-2 Screening criteria for choosing an EOR method for an oil resource
(modified from Taber et al. 1997)............................................................12 Table 3-1 Reservoir simulation model and fluid properties......................................35 Table 3-2 Well placement in Model H10 with a layer thickness of 10 m..................38 Table 3-3 Well placement in Model H7 with a layer thickness of 7 m......................41 Table 3-4 Well placement in Model H5 with a layer thickness of 5 m......................43 Table 4-1 Reservoir simulation model and fluid properties......................................89 Table 4-2 Well placement in Model H10-2Inj with a layer thickness of 10 m...........92 Table 4-3 Well placement in Model H7-2Inj with a layer thickness of 7 m...............95 Table 4-4 Well placement in Model H5-2Inj with a layer thickness of 5 m...............97 Table 4-5 Cross-sectional reservoir view and well configuration for best case
scenarios in dual injector-single producer..............................................121 Table 4-6 Cross-sectional reservoir view and well configuration for best case
scenarios in single injector-single producer and dual injector-single producer models.....................................................................................125
ix
List of Figures
Figure 2-1 Schematic representation of oil sand.........................................................8 Figure 2-2 A cartoon representation of asphaltene molecules: (A) the continental
and (B) the archipelago.............................................................................10 Figure 2-3 Schematic representation of Cyclic Steam Stimulation in a vertical well
configuration. ...........................................................................................14 Figure 2-4 Schematic representation of steam flooding technology.........................15 Figure 2-5 Schematic representation of In-Situ Combustion method.......................16 Figure 2-6 Cross-sectional view of the Steam-Assisted Gravity Drainage recovery
process......................................................................................................18 Figure 3-1 Cumulative steam oil ratio versus operation time for Model H10, (A)
aligned (B) not aligned, (C) best case scenarios........................................48 Figure 3-2 Cumulative produced oil versus operation time for Model H10, (A)
aligned (B) not aligned, (C) best case scenarios........................................51 Figure 3-3 Oil recovery factor versus pore volume steam injected (PVSI) for
Model H10, (A) aligned (B) not aligned, (C) best case scenarios...............53 Figure 3-4 Heat loss versus time for Model H10, (A) aligned (B) not aligned,
(C) best case scenarios..............................................................................56 Figure 3-5 Cumulative steam oil ratio (cSOR) versus operation time for Model H7
(A) aligned (B) not aligned, (C) best case scenarios..................................59 Figure 3-6 Cumulative produced oil versus operation time for Model H7,
(A) aligned (B) not aligned, (C) best case scenarios..................................61 Figure 3-7 Oil recovery factor versus pore volume steam injected (PVSI) for
Model H7, (A) aligned (B) not aligned, (C) best case scenarios................63 Figure 3-8 Heat loss versus time for Model H7, (A) aligned (B) not aligned,
(C) best case scenarios..............................................................................65 Figure 3-9 Cumulative steam oil ratio (SOR) versus operation time for Model H5,
(A) aligned (B) not aligned, (C) best case scenarios..................................67 Figure 3-10 Cumulative produced oil versus operation time for Model H5,
(A) aligned (B) not aligned, (C) best case scenarios..................................70 Figure 3-11 Oil recovery factor versus pore volume steam injected (PVSI) for
Model H5, (A) aligned (B) not aligned, (C) best case scenarios................72 Figure 3-12 Heat loss versus time for Model H5, (A) aligned (B) not aligned,
(C) best case scenarios..............................................................................74 Figure 3-13 Steam oil ratio (SOR) versus time for best case scenarios........................76 Figure 3-14 Cumulative produced oil versus operation time for best case Scenarios.78 Figure 3-15 Oil recovery factor versus pore volume steam injected (PVSI) for best
case scenarios ..........................................................................................79 Figure 3-16 Heat loss versus time for best case scenarios...........................................80 Figure 3-17 Heat loss behaviour for best case scenarios.............................................81 Figure 3-18 Wellpair arrangements and temperature distribution for best case
scenarios. (A) H10-6, (B) H7-4, (C) H5-5....................................................82 Figure 4-1 Cumulative steam-to-oil ratio versus time for Model H10-2Inj cases ....101
x
Figure 4-2 Cumulative produced oil versus time for Model H10-2Inj cases ............102 Figure 4-3 Oil recovery factor versus pore volume steam injected for
Model H10-2Inj cases..............................................................................104 Figure 4-4 Heat loss versus time for Model H10-2Inj cases.....................................106 Figure 4-5 Cumulative steam-to-oil ratio versus time for Model H7-2Inj cases......109 Figure 4-6 Cumulative produced oil versus time for Model H7-2Inj cases..............111 Figure 4-7 Oil recovery factor versus pore volume steam injected (PVSI) for
Model H7-2Inj cases...............................................................................112 Figure 4-8 Heat loss versus time for Model H7-2Inj cases.......................................113 Figure 4-9 Cumulative steam-to-oil ratio versus time for Model H5-2Inj cases......116 Figure 4-10 Cumulative produced oil versus time for Model H5-2Inj cases..............117 Figure 4-11 Oil recovery factor versus pore volume steam injected (PVSI) for
Model H5-2Inj cases................................................................................119 Figure 4-12 Heat loss versus time for Model H5-2Inj cases.......................................120 Figure 4-13 cSOR versus time....................................................................................122 Figure 4-14 Cumulative produced oil versus time.....................................................122 Figure 4-15 Oil recovery versus PVSI..........................................................................122 Figure 4-16 Heat loss versus time..............................................................................122 Figure 4-17 cSOR versus time....................................................................................124 Figure 4-18 Cumulative produced oil versus time.....................................................124 Figure 4-19 Oil recovery versus PVSI.........................................................................124 Figure 4-20 Heat loss versus time..............................................................................124 Figure 4-21 cSOR versus time....................................................................................127 Figure 4-22 Cumulative produced oil versus time.....................................................127 Figure 4-23 Oil recovery versus PVSI.........................................................................127 Figure 4-24 Heat loss versus time............................................................................127
xi
List of Symbols, Abbreviations and Nomenclature Symbols Definition
SGoil Specific gravity of the oil Q Oil production rate K Permeability g gravity acceleration ρ Oil density ϴ Inclined angle of the steam interface from horizon µ Oil viscosity ξ Distance from the interface T Temperature Ts Steam temperature Tr Initial reservoir temperature U Velocity of the advancing front Α Reservoir thermal diffusivity Ø Porosity ΔSo (Initial – residual) oil saturation g Gravity acceleration h Reservoir net pay m dimensionless constant νs Oil kinematic viscosity
Abbreviations Definition
API American Petroleum Institute cSOR cumulative Steam-to-Oil Ratio CSS Cyclic Steam Stimulation DWS Downhole Water Sink EOR Enhanced Oil Recovery ES-SAGD Expanding Solvent-Steam Assisted Gravity Drainage HWF Hot Water Flooding ISC In-Situ Combustion OOIP Original Oil In Place SAGD Steam Assisted Gravity Drainage SAGP Steam And Gas Push SF Steam Flooding SOR Steam-to-Oil Ratio VAPEX Vapor Extraction HW Horizontal Well
1
CHAPTER 1. INTRODUCTION
SAGD is an effective commercial process for viscous oil recovery from oil sands
reservoirs with pay zone thickness greater than approximately 15 m (Gates, 2010).
However, in thinner oil sands reservoir (< 10 m), heat losses from the steam chamber to
the overburden and understrata are significant and therefore oil recovery will be
achieved at the cost of higher energy consumption or cumulative steam-to-oil ratio
(cSOR) compared to that of thicker reservoirs. This also implies that CO2 emissions per
unit volume oil produced will likely be higher in thinner reservoir than would be the case
from thicker reservoirs. The importance of thin oil sand reservoirs lies in the fact that
about 80% of oil sands resources exist in reservoirs with a net pay zone of less than 5 m
in Western Canada (Adams, 1982). Thus, there is a need for new efficient processes to
produce these resources.
Documented research on the design of recovery processes for thin oil sands reservoirs is
scarce. Zhao et al. (2014) discussed different thermal recovery strategies to produce
from thin (< 5 m) heavy oil reservoirs. They investigated four production methods
consisting of cold production without sand, alternating injection/production well steam
and hot water, steam flooding, and SAGD. They found that first and second processes
are not suitable due to high energy to oil ratio and relatively low recovery factor. Both
steam flooding and SAGD are applicable but they still suffer from large steam use. The
steam cost or cSOR can be reduced by using solvent as was investigated by Gates (2010)
2
and Zhao et al. (2013). These studies focused on heavy oil reservoirs where the in situ
viscosity of the oil is of order of thousands to a few tens of thousands of centipoise.
The focus of the research documented in this thesis is on the application of SAGD in thin
oil sands reservoirs (≤ 10 m) where the viscosity of the oil at original reservoir conditions
is of order of 1 million cP. Since the cost of steam is the major expense of SAGD
operation and main contributor to carbon dioxide emissions from the process, our aim
is the reduction of cSOR through proposing various well configurations to achieve higher
oil production.
1.1. Statement of the Problem
In spite large steam consumption, steam flooding and SAGD are potentially applicable
processes in thin oil sands reservoirs (Zhao et al., 2014). However, reduction of high
steam use per unit oil recovered poses a challenge. It is known that reservoir
parameters as well as operational parameters influence oil recovery and SAGD
performance. For example, decreasing the oil column thickness raises heat losses and
therefore the cSOR increases. In comparison with conventional SAGD, well pair
configuration can be changed through vertical and horizontal positioning as well as
changing the ratio of injector to producer wells. There is potential that different well
configurations may lead to delayed production or faster steam breakthrough thus
ultimately changing and hopefully improving the SOR. There is a need for a detailed
3
study to address these parameters when all other operational and reservoir parameters
kept constant. The main research question being investigated, within the context of
SAGD-like processes, here is how can well configuration be altered to improve the
efficiency and recovery factor in reservoirs that are thinner than 10 m?
1.2. Objectives of the Thesis
The main objective of this thesis is to study the performance of SAGD operation in thin
oil sands reservoir with a thickness of (≤ 10 m). The key issue, as reported in the
literature (Gates 2008; Gates 2010), suggests that the major challenge faced by thermal
(steam) recovery processes for thin reservoirs are heat losses to the overburden and
understrata – that is, the energy efficiency of the process is the key challenge. However,
if the residence time of steam is relatively short in the reservoir, there is potential that
the energy efficiency of the process could be improved by altering the well
configuration including spacing between wells and their vertical locations.
1.3. Research Methodology
The research question has been answered by using detailed thermal reservoir
simulation. In a thermal reservoir simulator, the material and energy balances are
solved by using a numerical method – in this case, the domain is tessellated into grid
4
blocks and the finite volume method is used. In the research documented here, the
commercial thermal reservoir CMG STARSTM was used (CMG, 2013).
In total, 59 cases were studied each having different well configurations in three
reservoir thicknesses: 5, 7, and 10 m. Specifically, 11 single injector-producer well pairs
and 10 double injector-single producer well pairs were evaluated in the case where the
reservoir thickness was equal to 10 m (these cases are referred to as Model H10).
Similarly, in the 7 m reservoir model (referred to as Model H7), 8 numbers of single
injector-producer well pairs and 8 number of double injector-single producer well pairs
were tested. Finally, in the 5 m reservoir model (Model H5), 13 numbers of single
injector-producer well pairs and 9 number of double injector-single producer well pairs
were simulated. For each model, the performance of the recovery process was
evaluated on the basis of its cumulative steam oil ratio, cumulative oil production,
cumulative heat loss, and oil recovery factor.
1.4. Outline of the Thesis
Chapter 1 presents an overview on the thesis including a general introduction,
statement of the problem, objectives, research methodology and the outline.
Chapter 2 provides a general literature review on oil sands and in particular, recovery of
thin oil sands reservoirs. This includes oil sands resources, composition and chemistry.
5
Enhanced Oil Recovery (EOR) methodologies are discussed with a focus on thermal
recoveries. As SAGD is the method of choice for thicker oil sands reservoirs, in this
research work, it is briefly reviewed in Chapter 2 (process description, Butler’s theory,
variants of SAGD, SAGD challenges, and SAGD performance).
Chapters 3 and 4 contain the main research work. Both chapters have been submitted
as two manuscripts to peer-reviewed journals.
Chapter 3 describes the performance of Steam Assisted Gravity Drainage in thin oil sand
reservoirs using single injector-single producer well pair configuration. Specifically, the
influence of the injection and production well pair configuration in a homogeneous
formation with thicknesses of 5, 7, and 10 m was investigated. The wellpairs were
relocated to make different patterns where the spacing between the injection and
production wells were changed both horizontally and vertically. Also their locations with
respect to the overburden and understrata rock were varied.
Chapter 4 presents an analysis of the performance of Steam Assisted Gravity Drainage in
thin oil sand reservoirs using single producer-double injector well arrangement. The
results of this study are compared with the results of best case scenarios in Chapter 3.
Chapter 5 lists concluding remarks and recommendations for further research.
6
CHAPTER 2. LITERATURE REVIEW
2.1. Oil Sands Resources
The oil sands deposits hosted in Western Canada are the third largest proven deposit of
crude oil in the world (Natural Resources of Canada, 2013). Canada, along with
Venezuela holds 90% of the world’s heavy oil and bitumen (Nasr, 2005). The largest
reserves of bitumen are located in the Athabasca, Cold Lake and Peace River oil sands
deposits in Alberta with the average deposit depth of 300, 400, and 500 m, respectively
(Nasr, 2005).
The Athabasca oil sands were first explored by Pond in 1778 and a geological survey was
initiated in 1875 (Govier, 1965). The first wells were drilled between 1897 and 1898 at
Pelican Rapids on the Athabasca River (Govier, 1965). According to the Alberta
Department of Energy (ADOE, 2014), 167.2 billion barrels of proven reserves in the oil
sands deposit exist in Northern Alberta. Nearly 80% are recoverable through in-situ
recovery processes whereas the remainder is shallow enough to be recovered by
surface mining. One of the key uncertainties for oil sands reservoirs is the choice of the
recovery process that will yield the greatest recovery factor and rates at target water
consumption and greenhouse gas emissions. The two most used recovery processes for
oil sands reservoirs are the Cyclic Steam Stimulation (CSS) and Steam-Assisted Gravity
7
Drainage (SAGD) processes (Gates, 2013). Each one of these processes uses different
recovery mechanisms and as a result, the choice of the recovery process itself is a factor
that affects recovery from the reservoir. For example, steam assisted gravity drainage
leads to about 50-60% recovery in comparison with cyclic steam stimulation which reach
to around 35-40% (ADOE, 2014).
Crude oil is classified based on the API gravity, defined by 141.5/SGoil – 131.5 (SGoil is the
specific gravity of the oil), in four major types known as light, medium, heavy, and extra
heavy. The lower the API gravity, the higher the oil viscosity. Table 2-1 lists different
types of crude oil based on API.
Table 2-1 API gravity classification of petroleum oil (API, 2013).
Classification API gravity Viscosity, cP
Light > 31.1 < 10
Medium 22.3 - 31.1 10 - 100
Heavy 10 - 22.3 100 - 10,000
Extra Heavy < 10 > 10,000
Bitumen is categorized as extra heavy oil and has a typical API gravity of 5 to 9 with a
viscosity of > 10,000 cP at room temperature (Speight, 2007). Oil sands are a mixture of
sand, water, clay and bitumen (Kleindienst, 2005). For a typical McMurray Formation oil
8
sands reservoir, the oil sand contains about 83% sand, 14% bitumen and 3% water (Nasr
and Ayodele, 2005). As stated by Speight (1978), a typical composition for Athabasca
bitumen is 84% Carbon, 10% Hydrogen, 0.9% Oxygen, O.4% Nitrogen and 4.7% Sulphur.
Fig. 2-1 illustrates a typical composition of oil sands.
Figure 2-1 Schematic representation of oil sand.
Observations of oil sands grains suggest that the particles are water wet and thus they
are coated by water film which may hold clay particles (Cottrell, 1963) as shown in Fig.
2-1. Bitumen sits within the pore space between the water films. The viscosity of
bitumen depends on temperature and drops significantly as the temperature raises.
Raicar and Proctor (1984) investigated the relationship between viscosity and
temperature for light crude oil, heavy oil, and several bitumens. They showed that the
initial viscosity (at about 10 ⁰C) is highest for Athabasca (106 cP), followed by Cold Lake
9
(105 cP) and Lloydminster (103 cP) deposits and it drops to around 100, 80 and 8 cP,
respectively, at elevated temperature of about 300 ⁰C.
2.2. Chemistry of Heavy Oil and Bitumen
The hydrocarbon components of petroleum can be divided in three classes known as
paraffins (saturated hydrocarbons, no ring structures), naphthenes (saturated
hydrocarbons with one or more rings), and aromatics (hydrocarbons with one or more
aromatic rings) (Speight, 2007). Bitumen is a complicated mixture of hydrocarbons,
consisting of 105 to 106 different molecules (Wiehe, 1999). It is very difficult to
characterize all the components. However, it is possible to classify them in groups of
compounds by utilizing different techniques such as distillation, solubility/insolubility,
and adsorption/desorption (Wiehe, 1999). The most common standard method
applicable to bitumen or heavy oil is to separate them in four general solubility groups
known as saturates, aromatics, resins, and asphaltenes (ASTM, 1999).
The viscosity of petroleum is significantly influenced by the presence and concentration
of asphaltenes. Asphaltene is defined as a fraction of petroleum that is not soluble in
paraffinic solvents, e.g. heptane, but soluble in aromatic solvents. The molecular
structure of bitumen and asphaltenes consists mostly of C-C, C-H, C=C bonds (in the
aromatic rings) and to a lesser degree C-S, C-O, C-N, S-H, and O-H covalent bonds. The
10
metal impurities are mostly attached to nitrogen in porphyrin and non-porphyrin
structures (Rahimi and Thomas, 2006). As a result molecular aggregation occurs.
Generally, two structural models are suggested for asphaltene (Rahimi and Thomas,
2006; Schulze et al., 2015) known as continental and island (archipelago) models.
Representations of these two models are illustrated in Fig. 2-2. In reality the asphaltene
is a combination of these two structures with different percentages. In the continental
model, asphaltenes are composed of large aromatic cores that also contain hetero-
aromatics and metallo-porphyrins. The aromatic cores are surrounded by functional
groups, including alkyl groups and alkyl carboxylic acids (Mullins et al., 2012). In the
archipelago model, smaller aromatic islands are joined together by alkyl chains (Gray et
al., 2011).
Figure 2-2 A cartoon representation of asphaltene molecules: (A) the continental and (B) the archipelago.
11
Due to their high viscosity, heavy oil and bitumen recovery processes require the
application of enhanced oil recovery methods. Under natural conditions, the oil is too
viscous to flow under primary drive mechanisms to the surface.
2.3. EOR Methodologies
Generally, to enable oil production after primary and secondary recovery processes,
enhanced oil recovery methods are applied to a reservoir. In typical practice, EOR is
performed via injection of a fluid into the reservoir to displace the remaining oil in the
reservoir. Displacement can be immiscible or miscible depending on the material
injected into the reservoir. In miscible displacement, the interfacial tension between the
injectant and oil is equal to zero. In immiscible displacement, the injected phase
displaces the oil phase from the reservoir. The overall displacement efficiency is related
to wettability, capillary pressure, interfacial and surface tension forces, and relative
permeability as well as the reservoir heterogeneity and oil to injectant mobility ratio
(Terry, 2001). Another key factor that influences the efficiency of oil displacement from
the reservoir is the physical arrangement of injection and production wells.
Enhanced oil recovery methods can be categorized as miscible, chemical, thermal and
microbial flooding processes (Terry, 2001). This classification is based on the main
mechanism of oil displacement and formation lithology (Kokal and Al-kaabi, 2010).
These oil mobilization mechanisms are known as oil viscosity reduction, solvent oil
extraction, and alteration of wettability. Miscible EOR is applicable to light oil reservoirs
12
through gas injection (e.g. carbon dioxide). Chemical EOR is based on mobility control by
adding polymers to reduce the mobility of the injected water and/or reduction of
interfacial tension through addition of surfactants, and/or alkali. Thermal EOR is
generally suitable for heavy oil and oil sands recovery. Thermal energy increases the oil
temperature leading to oil viscosity reduction in the reservoir. The main thermal EOR
strategies include in-situ combustion and steam injection e.g. steam flooding, steam-
assisted gravity drainage, and cyclic steam stimulation.
The choice of an EOR method depends on various reservoir properties such as depth,
and composition (Gates, 2013). Some of the screening criteria for choice of suitable EOR
methods are summarized in Table 2-2 (Taber, 1997). According to Table 2-2, the method
of choice for highly viscous and permeable reservoir is the application of thermal energy
utilizing steam.
Table 2-2 Screening criteria for choosing an EOR method for an oil resource (modified from Taber et al. 1997).
Method Gravity (°API)
Viscosity (cP)
Lithology Net thickness
(ft)
Average permeability
(md)
Depth (ft)
Immiscible gases
> 12 < 600 Not critical Not critical
Not critical > 1,800
Polymer > 15 < 150 Sandstone preferred
Not critical
> 10 < 9,000
Combustion > 10 < 5,000 Highly porous sandstone
> 10 > 50 < 11,500
Steam > 8 < 200,000 Highly porous sandstone
> 20 > 200 < 4500
Surface mining
> 7 Mineable oil sand
> 10 Not critical > 3:1 Overburden/sand
13
Generally, the main aim in recovery of extra heavy oil (bitumen) is the reduction of
viscosity to make it mobile. Thermal methods, as well as solvent aided methods or their
combinations, are considered as good choices. Thermal methods are usually used for oil
recovery in heavy and extra heavy oils (Farouq-Ali, 2003). The main technologies based
on thermal recovery are Cyclic Steam Stimulation (CSS), Steam Flooding (SF), In-Situ
Combustion (ISC), Steam-Assisted Gravity Drainage (SAGD), and Hot Water Flooding
(HWF). The choice of these thermal methods is dependent on the reservoir
characteristics. For example, Mukhametshina et al. (2014) evaluated the recovery
characteristics of bitumen (8.8 API, 53,000 cP at 21 ⁰C) through application of four
thermal recovery methods including SAGD, HWI, SF and ISC. At their specific reservoir
condition, ISC showed the highest recovery factor and HWI the lowest recovery. SAGD
came second but at higher energy cost. SF showed similar results as WF but again at
higher energy consumption. They suggested that a hybrid method consist of HWI and
ISC works best for their particular reservoir.
2.3.1. Cyclic Steam Stimulation (CSS)
CSS method works based on the injection of steam at high pressure and temperature
into a reservoir. In the first step, steam is injected over a period of time into the
reservoir. In some operations, this is done above the fracture pressure and thus steam
fracturing is done in the reservoir (Gates, 2013). In other cases, steam is injected under
the fracture pressure. After the steam injection period is done, the well is shut in and
14
the hot steam zone in the reservoir further distributes its heat to oil sand there – this is
referred to as the soak period. The heated bitumen now has lower viscosity than its
original value, typically in the hottest zones of the reservoir equal to less than 20 cP.
The soak period may take up to 2 weeks. After the soak period is complete, the well is
re-opened for production and reservoir fluids are produced from the reservoir. After
the production rate of oil has dropped to a threshold value, the well is shut in and the
cycle starts again with steam injection. Typical recovery factors for CSS range from 20%
to 40% of the original oil in place (OOIP) with steam-to-oil ratios between 3 and 5
(Gates, 2013; Santos et al., 2014). A schematic representation of three steps in CSS
technology including injection, soak and production is illustrated in Fig. 2-3 in a vertical
well configuration.
Figure 2-3 Schematic representation of Cyclic Steam Stimulation in a vertical well configuration.
15
2.3.2. Steam Flooding (SF)
SF is based on continuous injection of high-pressure steam through a vertical injector
into a reservoir to create a hot zone which moves continuously across the reservoir
displacing oil to production wells. The latent heat of steam is transferred into the oil
zone and decreases the viscosity and thus raises the oil mobility. Steam zone is
expanded and the mobile oil is derived towards a vertical producer. In typical practice,
the oil recovery is about 50% of OOIP (Matthews, 1983; Nasr, 2005). A schematic
representation of SF technology is illustrated in Fig. 2-4.
Figure 2-4 Schematic representation of steam flooding technology.
16
2.3.3. In-Situ Combustion (ISC)
The process was patented in 1923 in USA and it is known as the oldest thermal recovery
method (Breston, 1958). A schematic representation of ISC technology is depicted in Fig.
2-5.
In-situ combustion works based on the oxidation of a small fraction of the reservoir oil.
Then, the combustion zone heats the oil and generates gas that displace the oil towards
the production well (Breston, 1958; Kendall, 2009). Although ISC appears to be an
effective recovery method for conventional as well as bitumen and heavy oil reservoirs
(Dayal et al., 2010), there have been no strong success cases in field operations.
Figure 2-5 Schematic representation of In-Situ Combustion method.
17
In bitumen and heavy oil application, the combustion front must be kept at high
temperature to provide the heat to keep the oil mobilized (Alamatsaz et al., 2011). This
implies that the Air-to-Oil Ratio (AOR) and the injection pressure are critical parameters
for process operation.
2.4. SAGD Process
Butler et al. (1981) combined the idea of bitumen mobilized by steam injection and
gravity drainage with the horizontal wellpair concept for the first time in Alberta in
1979. According to Edmunds and Chhina (2001), the Steam-Assisted Gravity Drainage
(SAGD) concept was used at an earlier time in steam flooding process using vertical
wellpairs in California (Doscher, 1966). The gravity drainage concept is depicted in Fig.
2-6. The main idea behind this method is rising of the injected steam at the bottom of
reservoir to heat up and decrease the oil viscosity. As a result the mobile oil and steam
condensate falls due to gravity and are collected simultaneously at the lower production
well.
In SAGD, a horizontal injection well and parallel horizontal production well is drilled near
the bottom of the reservoir with a certain vertical distance (e.g. 5 m) between them.
Prior to steam injection, the well pair is heated up by steam circulation to establish a
thermal communication between them. Then steam is introduced to the reservoir
18
through upper well and the mobile heavy oil and condensate are produced from the
reservoir through the lower well. A steam chamber develops in the chamber as oil is
drained from it. In the vertical direction, the steam chamber expansion rate is rapid until
it reaches the overburden cap rock. Thereafter, the steam chamber expands sideways
and downwards in the reservoir. At the edge of the chamber, steam loses its latent heat
to the oil sand and the bitumen beyond the edge of the chamber is heated via thermal
conduction.
Figure 2-6 Cross-sectional view of the Steam-Assisted Gravity Drainage recovery process.
19
The size of steam chamber and its uniform growth depend on how well the steam is
distributed within the reservoir and how uniform heat transfer is occurring at the edge
of the steam chamber. Permeability heterogeneity (Gotawala et al., 2010) and wellbore
undulation (Shen, 2011) can lead to a non-uniform steam chamber development in
SAGD.
Heat transfer is a vital element of SAGD. Various studies have been done to pin down
the importance and dominancy of convective and conductive heat transfer at the
interface or edge of SAGD steam chamber. Butler and Stephens (1981) considered
thermal conduction as the main heat transfer mechanism in SAGD operation and
regarded thermal convection as negligible. Further studies performed by Reis (1992),
Liang (2005) and Nukhaev et al. (2006) supported that idea. Farouq-AIi (2005) drew the
attention to the large volume of flowing steam condensate and expressed thermal
convection as dominant heat transfer mechanism. Ito and Suzuki (1996) using numerical
simulation showed that convection is dominant as well. Edmunds (1999) and Ito (1999)
suggested that the ratio of thermal convection to conduction is either less than 5%
(Edmunds, 1999) or around 55% (Ito, 1999), respectively. Sharma and Gates (2011)
showed that thermal convection provides a contribution to heat transfer at the edge of
the steam chamber. However, the increase in the heat-transfer rate by convection may
not necessarily translate to higher oil rates. They explained this behavior by relative
permeability effects at the chamber edge. Irani and Ghannadi (2013) investigated the
relative role of thermal convection in heat transfer through development of a
20
mathematical model by including both convection and conduction heat transfer to solve
the energy balance and pressure-driven condensate flow normal at the edge of SAGD
steam chamber. They concluded that convection can transfer a relatively large amount
of heat at the edge of steam chamber. However, it cannot be translated to high
temperature enhancement and supported the assumption of conduction-dominated
heat transfer. Irani and Gates (2013) spread more light on the subject of heat transfer in
SAGD process and investigated the relative roles of convective heat flux both parallel
and normal to the edge of the steam chamber. They suggested that the convective heat
flux associated with flow parallel to the chamber edge is minor compared with that
normal to the edge.
2.4.1. SAGD Analytical Model (Butler’s Theory)
Butler et al (1981) derived the oil production (drainage) rate equation for the SAGD
process based on the assumption of conductive heat transfer and Darcy’s Law. The final
result was:
𝑞𝑞 = 2�2∅∆𝑆𝑆𝑜𝑜𝑘𝑘𝑘𝑘𝑘𝑘ℎ
𝑚𝑚𝑣𝑣𝑠𝑠
where
21
q = Oil production rate
Ø = Porosity
ΔSo = Difference between Initial and residual oil saturation
K = Permeability
g = Gravity acceleration
α = Reservoir thermal diffusivity
h = Reservoir net pay
m = dimensionless constant, varies between 3 to 5 and depends on the oil
viscosity-temperature relationship
νs = kinematic oil viscosity
The temperature profile beyond the edge of the chamber used to derive this result is as
follows:
𝑇𝑇 − 𝑇𝑇𝑠𝑠𝑇𝑇𝑠𝑠 − 𝑇𝑇𝑟𝑟
= 𝑒𝑒−𝑈𝑈𝑈𝑈𝛼𝛼
where
T = Temperature
Ts = Steam temperature
Tr = Initial reservoir temperature
U = Velocity of the advancing front
ξ = Distance from the interface
α = Reservoir thermal diffusivity
22
Butler’s theory reveals that the oil production rate depends on several reservoir
parameters including the porosity, permeability, net pay thickness, reservoir thermal
diffusivity, bitumen kinematic viscosity and initial oil saturation.
2.4.2. SAGD Variants
One of the drawbacks of SAGD is its high steam consumption and therefore high cost of
steam production as well as higher greenhouse gas emissions (from carbon dioxide
resulting from the combustion of fuel for steam generation). There are several
technologies that have been developed in an attempt to reduce the energy and
environmental intensities of SAGD. These include the Steam And Gas Push (SAGP,
Butler, 1998; Jiang et al., 2000; Ito et al 2001), Expanding Solvent SAGD (ES-SAGD,
Butler, 1998; Nasr, 2005; Gates and Chakrabarty, 2008), and Vapor Extraction (VAPEX,
Butler, 1998) processes. In these three methods, additives are co-injected into the
reservoir with steam. They all use the same well configuration as that of SAGD. These
processes will not be discussed here as they are not the focus of the research
documented in this thesis.
2.4.3. SAGD Performance
The success of SAGD performance depends on the reservoir properties as well as
operational parameters. Reservoir properties include porosity, thickness, gas saturation,
23
permeability, viscosity and API gravity, wettability, heterogeneity, lithology, and geo-
mechanics. Operational parameters include start-up procedure and steam quality,
length, spacing and, placement of horizontal wells, sub-cool temperature or steam trap
control, pressure, steam chamber monitoring and size estimation, and well bore design.
The influence of these parameters has been investigated by various researchers. For
example Albahlani and Babadagli (2008) conducted a review on the influence of both
operational and reservoir properties on SAGD performance.
Sasaki et al. (1999, 2001) performed an experimental investigation with laboratory scale
together with reservoir simulation to study the role of reservoir layer thickness, steam
injection pressure and vertical spacing between SAGD well pair. Das (2005) conducted a
study based on analytical and simulation results and discussed role of well bore design
and operating pressure on SAGD performance and oil recovery rate. He reported that
lower operational pressure causes more challenging lift processes. However, low
pressure works in favour of water treatment at later time due to lower H2S production
and reduction in the amount of dissolved silica.
Carlson (2003) investigated the role of geomechanics and its influence on SAGD
production. Geo-mechanics directly ties with properties such as sampling procedure
(e.g. coring), formation properties (e.g. permeability, porosity, bitumen, water and gas
saturations), shearing and dilation. Therefore, by changing the operational conditions it
is possible to use geo-mechanics in favour of SAGD performance.
24
Su et al. (2012) developed a detailed 3D point bar model to determine the impact of
heterogeneity on SAGD performance in the McMurray Formation in the Long Lake area.
Their results revealed that SAGD orientation within the heterogeneous point bar has an
influence on the performance of SAGD. However, it requires more investigation to find
out specific well pairs arrangement to achieve an optimized cSOR and oil recovery.
Wang and Leung (2015) investigated the effect of lean zones and shale distribution on
the performance of SAGD in typical Athabasca oil sands with a pay zone of 30 m. In
particular, they studied heterogeneous distribution of shale barriers and lean zones
through variation of location, continuity, size, saturation, and proportions.
2.5. Thin Oil Sands Reservoirs
In thin oil reservoirs, in the literature, this is any reservoir that has thickness less than
about 15 m. For example Adams (1982) defined the thickness as < 5 m, whereas Gurgel
et al. (2009) described it as 5 to 15 m, and Feng et al. (2014) as 4 to 10 m and so on. In
this research work we classified a thin oil sands reservoir as any oil sands reservoir with
thickness less than 10 m.
About 80% of oil sand resources exist in reservoirs with a net pay zone of less than 5 m
in Western Canada (Adams, 1982). Thus, there is a need for new efficient processes or
strategies to produce these resources. In spite of potentially large steam consumption,
25
steam flooding and SAGD are potentially applicable processes in thin oil sands reservoirs
(Zhao et al., 2014).
Doscher and El-Arabi (1983) studied a pilot steam injection process in thin oil sands
reservoir (about 5 to 7 m in thickness) in California. They concluded that a higher steam
injection rate at the beginning of the process leads to higher oil recovery due to faster
arrival of the oil bank to producer well. Feng et al. (2014) investigated the parameters
affecting the steam breakthrough in a steam flooding operation in thin layer ultra-heavy
oil reservoirs (viscosity ≥ 50,000 cP) with a thickness of 4 to 10 m. Their results implied
that the formation of a steam breakthrough channel depended on the reservoir
permeability and oil saturation. They suggested that injection of nitrogen foam at initial
stage of steam breakthrough can help the process by hindering the steam breakthrough.
Gates (2010) evaluated the operating conditions of ES-SAGD in thin heavy oil reservoirs
with a thickness of 8 m. Specifically, he used stochastic optimization to determine the
optimal injection pressure and solvent concentration in the injected steam. The results
of the study revealed that these two parameters have an impact on the system energy
efficiency. A comparison between SAGD and ES-SAGD showed that ES-SAGD leads to
lower steam and energy usage than that of SAGD. Furthermore, it was shown that the
performance of an optimized thermal-solvent added process is comparable to VAPEX. It
implies that the injected steam provides sufficient thermal energy to keep the area near
26
the wellpair hot. Thus, solvent remains longer in the vapor phase and leads to
promotion of the oil mobilization process.
Gurgel et al. (2009) studied the influence of operational parameters in steamflooding
process in thin oil reservoirs with a thickness of 5 to 15 m. They concluded that
horizontal permeability, water and oil zone thicknesses, and thermal conductivity have
an influence on cumulative oil production. A reservoir of 5 m height showed a better
response to optimization process (steam injection rate and well distances).
Chang (2013) investigated the application and economics of horizontal well (HW)-CSS
for thin oil sand and heavy oil reservoirs. It was shown that HW-CSS is not economical
for layer thicknesses of 5 to 8 m. While a thickness of 11 m with production duration of
8 years using 8 well pads will be economical.
2.6. Well Spacing and Configuration
Well configuration and spacing or the pattern of SAGD well pairs within the reservoir
can be defined in different ways to achieve different contributions from different drive
mechanisms. Usually, well spacing reflects the distance between SAGD well pairs and is
assumed as repeated within the overall pattern of the SAGD well pairs. The
configuration indicates the vertical and horizontal distances or offsets between the
SAGD wells in each repeat unit. The SAGD well pair configuration can include horizontal,
27
vertical or slanted well arrangements. Here a review is given on the studies that are
conducted on both well pair spacing and configurations and their subsequent influence
on SAGD performance.
Joshi (1986) investigated the SAGD performance through laboratory experiment by
comparing cSOR in vertical and horizontal well configuration. The results showed that
horizontal SAGD gives a better performance. Miller and Xiao (2007) proposed a well
configuration in which vertical wells are drilled in between classical SAGD well pair
spacing in a heavy oil reservoir with a pay thickness of 20 m to improve the production.
They stated through field observation and numerical simulation that vertical production
wells are able to produce the remaining oil not produced by horizontal producer wells
leading to increased oil recovery. Jimenz (2008) performed an analysis on the
performance of SAGD projects in Canada and found out that a inter well pair spacing of
100 m is the most common with the best results with respect to field performance and
that SAGD is mostly applicable to reservoirs with a net pay zone greater than 15 m.
Mojarab et al. (2011) proposed dipping-injector SAGD well configuration for application
in Athabasca and Cold Lake reservoirs with a pay zone thickness of 20 m. Their
simulation results revealed a better performance in comparison with conventional SAGD
through an improvement in thermal efficiency and growth of a more uniform steam
chamber.
28
Cheung (2013) investigated the influence of SAGD well spacing while considering central
processing facility constraints (steam supply and fluid processing capacity). She reported
that the optimal well spacing is around 85 to 125 m with the most economical distance
of 100 m. Verney (2015) investigated the role of well pair length and spacing on SAGD
production through assessment of cSOR, bitumen rate and recovery factor for 1,111
well pairs in the McMurray Formation. Well pair spacing and length were varied in the
range of 40 to 160 m and 400 to 1400 m, respectively. He concluded that well length
does not influence SAGD performance. However, tighter inter-well spacing lead to lower
cSOR and higher recovery factors. Gupta et al. (2015) investigated the impact of well
spacing on SAGD solvent aided processes (SAGD/SAP) using results from a field trial in a
net pay zone of 24 m. Their results confirmed that it is feasible to apply a wider spacing
in SAGD/SAP system in comparison with conventional SAGD. Thus, SAGD/SAP requires
less number of well pairs which in turn reduces the cost, footprint of surface facilities
and the environmental impact.
The well configuration in conventional SAGD consists of a parallel horizontal well pair
which is drilled 5 m apart. Different well configurations have been presented in the
literature to improve SAGD performance. However, there are not many studies
contributed to the bituminous thin oil reservoirs of thickness less than 10 m. Here,
several studies on the SAGD well configuration are reviewed.
29
Tamer and Gates (2012) investigated the impact of position and geometry of the
injector wells in a McMurray Formation reservoir model with properties drawn from the
Dover SAGD Phase B. The reservoir thickness is equal to 24 m. In this study, different
injector well configurations including typical SAGD, offset SAGD and vertical/horizontal
well combination were evaluated. They suggested that a number of vertical injectors
can deliver steam to the reservoir more efficiently than a single horizontal well at early
stages of the process. This is due to smaller exposure of steam chamber to the
overburden. Regarding offset SAGD, they found out that greater offset leads to the
creation of larger steam-chamber volume and therefore higher oil recovery. However,
the initiation of thermal communication between the injection and production wells at
the start of the process revealed to be both challenging and demanding of relatively
large volumes of steam.
Tavallali et al. (2011, 2012) investigated the impact of well configurations for SAGD in
Athabasca McMurray Formation with a net pay thickness of 20 m and in Lloydminster
heavy oil reserve with a net pay thickness of 10 m. The viscosity of the oil in the
Lloydminster reservoir was equal to 5000 cP at the reservoir temperature. According to
Tavallali et al., under certain circumstances, it is possible to increase well spacing
because the lower viscosity allows for establishment of easier thermal communication
between the well pairs in comparison with higher viscosity reservoirs such as those in
the McMurray Formation where the viscosity of the oil is typically above 1 million cP.
Different well configuration including conventional SAGD, vertical injector, reversed
30
horizontal injector, inclined injector, parallel inclined injectors and multi lateral
produced were proposed and studied via numerical simulation for Athabasca reservoirs.
They observed no advantage in using vertical injectors in Athabasca reservoirs – this
contradicts Tamer and Gates’ results and results obtained from field operations (Miller
and Xiao, 2007). The best result was obtained with the application of reversed
horizontal steam injectors. Different well configurations including conventional SAGD,
offset injector, multi lateral producer and ZIGZAC pattern were proposed for thinner
Lloydminster reservoir (Tavallali et al., 2012). Their results revealed that a maximum
offset distance of 12 m leads larger drainage volume but at higher cost of cSOR. The
multi lateral configuration showed the most optimum results with a cSOR of about 5
m3/m3.
Among the conducted studies on the SAGD well pair configuration, it was found that
there are few dedicated to the role of vertical well pair spacing on SAGD performance
especially in case of thin oil sands reservoirs of ≤ 10 m. Sasaki et al. (1999), based on
their laboratory experimental results, concluded that oil production rate increases with
increasing vertical spacing in a conventional SAGD well pair configuration. However, it
comes with the cost of longer lead time for gravity drainage to initiate oil production. In
another study, Sasaki et al. (2001) suggested that the decrease in vertical spacing causes
faster establishment of thermal communication and an increase in spreading rate of
steam and leads to higher amount of oil production.
31
Tavallali et al. (2011) investigated the impact of vertical well distance on SAGD
performance in Athabasca McMurray Formation with a net pay thickness of 20 m. It was
shown that the preheating period is shortened when the distance is less that 5 m
(conventional SAGD) with no significant effect on SAGD performance. The performance
was decreased with increasing the vertical distance. They reported that a distance
within range of 3 to 6 m is desirable with an optimum distance of 4 m.
2.7. What is Missing in the Literature?
The literature review reveals that a large number of studies of the SAGD process has
been conducted to improve its performance by altering the operating strategy and the
well configuration. However, there are none that investigate how well configuration
can be modified to improve the performance of SAGD in thin (less than 10 m) oil sands
reservoirs (reservoirs with oil viscosity of order of a million cP). The research
documented in this thesis fills this gap.
32
CHAPTER 3. PERFORMANCE OF STEAM ASSISTED GRAVITY DRAINAGE IN THIN OIL SAND RESERVOIRS: WELL
CONFIGURATION
Summary
The performance of Steam Assisted Gravity Drainage (SAGD) is studied in thin oil sand
reservoirs. Specifically, the influence of the injection and production well pair
configuration in a homogeneous formation with thicknesses of 5, 7, and 10 m was
investigated. The well pairs were relocated to make different patterns where the
spacing between the injection and production wells were changed both horizontally and
vertically. Also their locations with respect to the overburden and understrata rock were
varied. SAGD performance was assessed numerically via a thermal reservoir simulator
and the cumulative steam oil ratio (cSOR), cumulative oil production, cumulative heat
loss, and oil recovery factor were compared.
3.1. Introduction
SAGD is an effective commercial process for viscous oil recovery from oil sands
reservoirs with a pay zone thickness of greater than approximately 15 m (Gates, 2010).
However, in thin oil sands reservoirs with thickness lower than 10 m, heat losses from
the steam chamber to the overburden and understrata are significant and therefore oil
33
recovery will be achieved at the cost of higher energy consumption or cumulative
steam-to-oil ratio (cSOR) compared to that of thicker reservoirs. This also implies that
carbon dioxide emissions per unit oil produced will likely be higher in thinner reservoir
than would be the case from thicker reservoirs. The importance of thin oil sand
reservoirs lies in the fact that about 80% of oil sand resources exist in reservoirs with a
net pay zone of less than 5 m in Western Canada (Adams, 1982). Thus, there is a need
for new efficient processes to produce these resources.
Zhao et al. (2014) discussed different thermal recovery strategies to produce from thin
(< 5 m) heavy oil reservoirs. They investigated four production methods consisting of
cold production without sand, alternating injection/production well steam and hot
water, steam flooding, and SAGD. They found that first and second processes are not
suitable due to high energy to oil ratio and relatively low recovery factor. Both steam
flooding and SAGD are applicable but they still suffer from large steam use. The steam
cost or cSOR can be reduced by using solvent as was investigated by Gates (2010) and
Zhao et al. (2013). These studies focused on heavy oil reservoirs where the in situ
viscosity of the oil is of order of thousands to a few tens of thousands of centipoise.
The focus of the research documented here is on the application of SAGD in thin oil
sands reservoir (thickness less than 10 m) where the impact of well configuration on
process performance will be investigated. In these reservoirs, the viscosity of the oil at
original reservoir conditions is of order of 1 million cP. Since the cost of steam is the
34
major expense of SAGD operation and main contributor to carbon dioxide emissions
from the process, our aim is the reduction of cSOR with higher oil production.
3.2. Reservoir Simulation Model
The reservoir simulation models used in the research documented here consist of a set
of two-dimensional homogenous model with horizontal well pairs. The reservoirs do
not have gas cap or bottom water zones. Specifically, three reservoir models were
developed with oil sand intervals of 5, 7, and 10 m thickness, respectively labeled as
Model H5, H7, and H10. A regular Cartesian grid system was used to discretize the
models with dimensions of 58 grids with a block size of 0.8 m in the cross well direction,
1 grid block with size of 750 m in the downwell direction. In the vertical direction, there
are 10, 14, and 20 grid blocks in the H5, H7, and H10 models, respectively, all with
dimensions equal to 0.5 m. The reservoir model properties and parameters are shown
in Table 3-1.
Simulations were performed utilizing a commercial thermal reservoir simulator, CMG
STARSTM Version 2013 (CMG, 2013). In this finite volume based thermal reservoir
simulator, the conservation of energy and mass equations are solved over each grid
block together with the phase behavior and relative permeability curves for the gas,
aqueous, and oil phases. At the lateral sides of the model, symmetry boundary
35
conditions were imposed (no flow or heat transfer). In other words, the width of the
reservoir represents the horizontal well spacing and implies that the well pairs are part
of a larger pattern.
Table 3-1 Reservoir simulation model and fluid properties.
Property
Value
Net pay, m 10, 7 and 5 SAGD wellpair length, m 750 Horizontal permeability, mD 4000 Vertical permeability, mD 2000 Average porosity 0.3 Initial oil saturation 0.75 Initial water saturation 0.25 Irreducible water saturation (Swr) 0.15 Residual oil saturation with respect to water 0.20 Relative permeability to oil at irreducible water 1.0 Relative permeability to water at residual oil 0.992 Residual gas saturation (Sgr) 0.005 Residual oil saturation with respect to gas 0.005 Relative permeability to gas at residual oil 1.0 Relative permeability to oil at critical gas Krw at irreducible oil (KRWIRO)
0.992 0.1
Residual oil for gas-liquid table endpoint saturation (SORG) 0.005 Initial temperature, °C 20 Initial pressure, kPa 2000 Rock heat capacity, J/m3 °C 2.600x106 Rock thermal conductivity, J/m day °C 6.600x105 Water phase thermal conductivity, J/m day °C 5.350x104 Oil phase thermal conductivity, J/m day °C 1.150x104 Gas phase thermal conductivity, J/m day °C 5.000x103 Bitumen Molecular weight, kg/kmol Critical temperature, °C Critical pressure, kPa Dead oil viscosity, cP at 10°C 100°C 200°C
465
903.85 792
1587285 203.91
9.71
36
Table 3-1 Reservoir simulation model and fluid properties (continued).
Liquid phase component viscosity (cP) versus temperature curves (methane viscosities are liquid equivalent viscosity) T (°C) µwater µoil µmethane
The change of heat loss is not significant compared to the other cases (Cases H5-3, 5, 6
and 7). The temperature profile showed a better distribution in comparison with H5-2
but at cost of higher heat loss.
In Fig. 3-12B, a change of the horizontal separation between the wells has profound
effects on SAGD performance since it delays oil production significantly.
According to Fig. 3-12C, the order with respect to heat losses from the highest to the
lowest is Cases H5-11, H5-10, H5-9 and H5-12. Similar with what observed in the
previous models, not vertically aligned cases lead to lowest heat loss.
Similar to previous models, the same trend in choosing the best cases scenarios (here:
H5-4, H5-5, H5-11 and H5-13) for cSOR, cumulative oil production and recovery factor
did not appear. A comparison between Cases H5-4, H5-5, H5-11 and H5-13 indicates
that, except for Case H5-11, the change in heat losses is not significant. The highest
heat loss is achieved when the wellpairs have a vertical distance of 3 to 4.5m from each
other. Introducing a horizontal separation reduces the heat loss as well as the oil
recovery factor.
76
3.4.4. Best Case Scenarios
The best case scenarios for Models H10, H7 and H5 are compared with each other’s and
a thicker reservoir layer of 25 m known as base case in the following sections.
3.4.4.1. Cumulative Steam-to-Oil Ratio
The cumulative steam-to-oil ratios (cSORs) for the best case scenarios and the thicker
base case are shown and compared with each other in Fig. 3-13. In all cases, lower cSOR
is achieved by positioning the producer well at 0.25 m above understrata with a vertical
alignment of the injection and production wells. The optimum vertical distance attained
for the 7, 10 and 25 m reservoirs was found to be equal to 5 m and 4.5 m for the 5 m
reservoir.
Figure 3-13 Steam oil ratio (SOR) versus time for best case scenarios.
0
1
2
3
4
5
6
7
8
0 100 200 300 400 500 600 700 800 900 1000
SOR
Cum
ulat
ive
, m3 /
m3
Time, Days
H25-Base Case H10-Case 6 H7-Case 4 H5-Case 5
77
A comparison between the cases reveals that the thicker the layer, the smaller the
cSOR. Also the cSOR profile, between 100 to 300 production days, shows a local
maximum of the cSOR. Beyond this peak, the cSOR drops to a local minimum and then
it grows once again for the best case scenarios. The early peak is related to higher
capacity of reservoir to accept steam (larger net pay zone) and increasing heat losses to
the rock outside the reservoir. Slower growth of the cSOR beyond the local minimum is
observed for the thicker reservoirs. The cSOR for the thicker base case leads to a gradual
decrease. Approximately, cSOR with a value of around 3.5, 4.2, 5.4 and 7.2 m3/m3 are
obtained for oil reservoirs of thickness 25, 10, 7 and 5 m, respectively.
3.4.4.2. Cumulative Produced Oil
The cumulative produced oil volume profiles for the best case scenarios together with
the thicker base case are shown in Fig. 3-14. The same trends of the rise of the
cumulative oil production for all oil reservoir thicknesses are observed. However, the
greater the thickness of the reservoir, the higher is the cumulative oil produced and the
oil rate. Obviously, a thicker layer consists of larger pay zone and consequently higher
amount of oil which allows for higher cumulative production at the same production
time.
78
Figure 3-14 Cumulative produced oil versus operation time for best case scenarios.
3.4.4.3. Oil Recovery Factor
The profiles of the oil recovery factor versus the pore volume injected steam (PVSI) for
the best case scenarios together with the thicker base case are presented in Fig. 3-15.
Even in the thinnest case, the recovery factor of 55% can be achieved but at higher cost
of injected steam (PVSI=2.8). The thicker the oil column, the higher the oil recovery
factor is given a PVSI. A range of 55 to 60% was observed for oil reservoir thickness of 5
to 10 m. For the base case of 25 m about 33% cumulative oil recovery observed at the
end of 900 operation days. Obviously due to the size of reservoir it takes longer time to
achieve higher recovery factor. For example, the cumulative recovery factor increased
0
10000
20000
30000
40000
50000
60000
70000
0 100 200 300 400 500 600 700 800 900 1000
Cum
ulat
ive
Oil
Prod
uctio
n , m
3
Time, Days
H25-Base Case H10-Case 6H7-Case 4 H5-Case 5
79
to 69% (H25-Base Case*) after 1500 operation days. To achieve the same recovery
factor, a greater PVSI is required as the oil reservoir gets thinner.
Figure 3-15 Oil recovery factor versus pore volume steam injected (PVSI) for best case scenarios.
3.4.4.4. Cumulative Heat Loss
The cumulative heat losses for the best case scenarios are depicted in Fig. 3-16. Ignoring
Case H5-12, it can be concluded that higher layer thicknesses lead to lower heat losses.
In all cases with the exception of H5-12, the producer well is located at 1.75 m above
the underburden and the injector well is positioned with a vertical and horizontal
distances of 2.5 and 8 m respectively. Turning our attention to Case H5-12, it is shown
that this particular well configuration leads to a lower heat loss than that of layer
0
10
20
30
40
50
60
70
0 0.5 1 1.5 2 2.5 3
Oil
Reco
very
Fac
tor
PVSI
H25-Base Case H25-Base Case* H10-Case 6
H7-Case 4 H5-Case 5
80
thickness of 7 and 10 m. Here both injector and producer wells are positioned at the
adjacent block to the underburden and 4 m apart. Evaluating SAGD well performance
based on the lowest heat losses does not make sense as far as we are interested in
achieving higher oil recovery and production.
Figure 3-16 Heat loss versus time for best case scenarios.
In our investigation, the best case scenarios from the heat loss point of view led to
poorest SAGD well performances.
The selected well configuration of best case scenarios (H10-6, H7-4 and H5-5) for cSOR,
cumulative oil production and oil recovery were similar and completely different from
the well configuration selected for heat loss best case scenarios (H10-9, H7-6 and H5-
12).
-1.5E+13
-1E+13
-5E+12
0
0 100 200 300 400 500 600 700 800 900 1000
Heat
Los
s Cu
m, J
Time, Days
H10- Case 9 H7-Case 6H5-Case 12 H5- Case 9
81
In order to be able to come in conclusion, we are selecting Cases H10-6, H7-4 and H5-5
and look at their heat loss behavior. This is depicted in Fig. 3-17 and reveals that layer
thickness of 10 m leads to lower heat loss (-1.67 x 1014 J). However, for the particular
well arrangement in H7-4 and H5-5 the heat losses almost are the same. Under the
same circumstances, the base case of 25 m thickness leads to a cumulative heat loss of
(-1.14 x1014 J).Roughly, thinner layers lead to 46% up to 59% more heat loss in
comparison with the thicker base case.
Figure 3-17 Heat loss behaviour for best case scenarios.
3.4.4.5. Temperature Distributions and Well Pairs Arrangement
The wellpairs arrangements as well as temperature distribution for best case scenarios
(regardless of best heat loss cases) are depicted in Fig. 3-18. As the oil reservoir
-2E+14
-1.5E+14
-1E+14
-5E+13
0
0 100 200 300 400 500 600 700 800 900 1000
Heat
Los
s Cu
m, J
Time, Days
H25-Base Case H10- Case 6 H7-Case 4 H5-Case 5
82
thickness decreases, the steam injector position gets closer to overburden. The injector
distance from the overburden for Cases H10-6, H7-4, and H5-5 are 4.75, 1.75 and 0.75
m, respectively. Therefore, a higher heat loss to overburden is expected. As the oil
reservoir thickness increases, better temperature distribution is expected due to smaller
heat loss and better heat transfer to the larger reservoir volume.
(A)
(B)
(C)
Figure 3-18 Wellpair arrangements and temperature distribution for best case scenarios. (A) H10-6, (B) H7-4, (C) H5-5.
83
3.5. Conclusions
The influence of well configuration on the production performance of SAGD operation
was studied in thin (less than 10 m thick) oil sand reservoirs. Production is possible with
oil recovery factors of 60% to 55% at cSORs equal to about 4, 5, and 7 m3/m3 for oil
reservoir thicknesses of 10, 7 and 5m, respectively. The best case scenarios were
achieved by positioning the producer well at 0.25 m above the understrata and having a
vertical wellpairs alignment with a distance of about 4 to 5 m. The steam injector
distance from overburden was varied between 0.75 and 4.75 m. The results imply that
with the decrease in oil reservoir thickness, position of the injector well gets closer to
the overburden. The cumulative heat losses were almost in the same range for oil
reservoir equal to 5 and 7 m and slightly lower for layer thickness of 10 m. In
comparison with a thicker layer reservoir of 25 m, higher heat losses were observed –
the increase in heat loss led to 46, 56 and 59% for layer thickness of 10, 7 and 5 m
respectively.
It can be concluded that vertical as well as horizontal distances between wellpairs, and
their position or distances in respect to over and underburden rock affect the SAGD
production. Finding an optimum value will have an influence in cSOR, heat loss, recovery
factor and oil production. Therefore, to maximize SAGD production in thin oil sands
reservoirs requires finding an optimized well configuration. Introducing a horizontal
separation between the injector and producer wells usually decreases SAGD
84
performance through delayed oil production. Under our experimental condition, a
horizontal offset of about 1.6 m was found to be optimal.
85
CHAPTER 4. PERFORMANCE OF STEAM ASSISTED GRAVITY DRAINAGE IN THIN OIL SAND RESERVOIRS: WELL PAIR
CONFIGURATION IN A SINGLE PRODUCER - DOUBLE INJECTOR SET UP
Summary
The performance of Steam Assisted Gravity Drainage (SAGD) is studied in thin oil sand
reservoirs where the basic unit of operation is a single producer with two steam
injectors. Specifically, the influence of the injection and production well triplet
configuration is investigated in homogeneous oil sands formations with thicknesses of 5,
7, and 10 m. Various well configurations were explored where the vertical and
horizontal spacing between the injection and production wells were varied. Thus, well
locations with respect to the overburden and understrata rock also varied. SAGD
performance was assessed numerically by using a thermal reservoir simulator and the
cumulative steam-to-oil ratio (cSOR), cumulative oil production, cumulative heat loss,
and oil recovery factor were compared. The results of study were compared with the
best case scenarios from single producer-single injector well configurations. The results
suggest that horizontal and vertical distances between injectors and the producer well,
their locations from the overburden and understrata and their vertical alignment impact
their performance. The results also show that the addition of an offset injector well
86
reduces cSOR under certain well configuration. Generally, the dual injector-single-
producer performs better than the single injector-single-producer.
4.1. Introduction
The Steam-Assisted Gravity Drainage (SAGD) oil sands recovery process is an effective
commercial process for viscous oil recovery when the reservoir thickness is greater than
about 15 m (Gates, 2010). However, in thinner oil sands reservoir (< 10 m), heat losses
from the steam chamber to the overburden and understrata are significant and the
process is considered inefficient and uneconomic. However, the importance of thin oil
sand reservoirs lies in the fact that about 80% of extra heavy oil resources exist in
reservoirs with a net pay zone of less than about 5 m in Western Canada (Adams, 1982).
Thus, there is a need for new efficient processes to produce these resources.
In Chapter 3, the importance and influence of well configuration on SAGD performance
in thin oil sand reservoirs (≤ 10 m) were investigated. The key parameters such as
vertical as well as horizontal separation between the injector and producer wells, and
their positions with respect to overburden and understrata were evaluated by using a
single injector-single producer system. It was shown that these parameters have an
influence on the cSOR, heat losses, recovery factor, and cumulative oil production.
Introducing a horizontal separation between the wells usually decreases SAGD
performance at the cost of delaying the onset of oil production. However, unless the
87
vertical distance is already less than the optimum value, then a small horizontal shift or
offset improves process performance. From the study in Chapter 3, a horizontal offset of
about 1.6 m between the wells is suggested.
The focus of the research documented here is an evaluation of a dual injector-single
producer well configuration in thin oil sands reservoir (≤ 10 m). In practice, the dual
injector could be completed in the reservoir using multilateral drilling technology and
thus, there are no technical barriers to using dual injection wells as envisioned in the
research documented here.
4.2. Reservoir Simulation Model
The reservoir simulation model consists of a two-dimensional homogenous model with
horizontal wellpairs. The reservoirs do not have gas cap or bottom water zones.
Specifically, similar to Chapter 3, three reservoir models were developed with oil sand
intervals of 5, 7, and 10 m thickness, respectively labeled as cases H5-2Inj, H7-2Inj, and
H10-2Inj. It should be noted that for the sake of comparison with Chapter 3 results, the
2Inj suffix added to model names implies that the SAGD well configuration consists of
single-producer-double injector wells. The second injector well is referred as 2Inj or
offset well here.
88
A regular Cartesian grid system was used to discretize the models with dimensions of
58 grids with a block size of 0.8 m in the cross well direction, 1 grid block with length
750 m in the downwell direction. In the vertical direction, there are 10, 14, and 20 grid
blocks in the H5-2Inj, H7-2Inj, and H10-2Inj models, respectively, all with grid block
dimensions equal to 0.5 m. The reservoir model properties and parameters are listed in
Table 4-1.
Simulations were performed using a commercial thermal reservoir simulator CMG
STARSTM version 2013 (CMG, 2013) which has been used in the oil sands industry for
over 15 years for SAGD simulation. In this finite volume thermal reservoir simulator, the
conservation of energy and mass equations are simultaneously solved over each grid
block together with the phase behavior and relative permeability curves for the gas,
aqueous, and oil phases. At the lateral sides of the model, symmetry boundary
conditions were imposed (no flow or heat transfer). In other words, the width of the
reservoir represents the horizontal well spacing and implies that the well triplets are
part of a larger pattern.
In our models, to establish thermal communication between injector and producer
wells, the steam circulation time was set equal to 3 months. In the model, this was
done by placing temporary heaters in the locations of the wells. When SAGD mode
started, the temporary heaters were turned off and steam injection and fluid production
started. The operation was simulated for up to 3 years.
89
Table 4-1 Reservoir simulation model and fluid properties.
Property
Value
Net pay, m 10, 7 and 5 SAGD wellpair length, m 750 Horizontal permeability, mD 4000 Vertical permeability, mD 2000 Average porosity 0.3 Initial oil saturation 0.75 Initial water saturation 0.25 Irreducible water saturation (Swr) 0.15 Residual oil saturation with respect to water 0.20 Relative permeability to oil at irreducible water 1.0 Relative permeability to water at residual oil 0.992 Residual gas saturation (Sgr) 0.005 Residual oil saturation with respect to gas 0.005 Relative permeability to gas at residual oil 1.0 Relative permeability to oil at critical gas Krw at irreducible oil (KRWIRO)
0.992 0.1
Residual oil for gas-liquid table endpoint saturation (SORG) 0.005 Initial temperature, °C 20 Initial pressure, kPa 2000 Rock heat capacity, J/m3 °C 2.600x106 Rock thermal conductivity, J/m day °C 6.600x105 Water phase thermal conductivity, J/m day °C 5.350x104 Oil phase thermal conductivity, J/m day °C 1.150x104 Gas phase thermal conductivity, J/m day °C 5.000x103 Bitumen Molecular weight, kg/kmol Critical temperature, °C Critical pressure, kPa Dead oil viscosity, cP at 10°C 100°C 200°C
465
903.85 792
1587285 203.91
9.71
90
Table 4-1 Reservoir simulation model and fluid properties (continued).
Liquid phase component viscosity (cP) versus temperature curves (methane viscosities are liquid equivalent viscosity) T (°C) µwater µoil µmethane
4.4.5. Best Cases for the Single Injector-Single Producer and Dual Injector-Single
Producer Models
The best cases for two major categories of single injector-single producer and dual
injector- single producer models are discussed here. Reservoir cross-sections and well
configurations for the best cases are depicted in Table 4-6.
The dual injector-single producer cases (not aligned) configurations led to best
performance whereas in the single injector-single producer cases, the aligned ones
appeared to perform better.
Table 4-6 Cross-sectional reservoir view and well configuration for best case scenarios in single injector-single producer and dual injector-single producer models.
Case Dual inj. - Single-prod.
Case Single-inj. - Single prod.
H10-2Inj-8
H10-6
H7-2Inj-8
H7-4
H5-2Inj-9
H5-5
Fig. 4-21 shows a comparison of cSOR performance. The dual injector-single producer
cases lead to lower cSOR values for all reservoir thicknesses. Having two injector wells in
126
comparison with one helped with faster growth of steam chamber and thus oil
mobilization.
Fig. 4-22 displays the cumulative produced oil as a function of time. The dual injector-
single producer cases deliver higher cumulative oil production.
Fig. 4-23 compares the oil recovery factor versus pore volume injected steam. The dual