University of Calgary PRISM: University of Calgary's Digital Repository Graduate Studies The Vault: Electronic Theses and Dissertations 2018-06-29 Optimum Switch Timing for Steam-Assisted Gravity Drainage (SAGD) after Cyclic Steam Stimulation (CSS) in a Viscous Oil Reservoir Liu, Siyuan Liu, S. (2018). Optimum Switch Timing for Steam-Assisted Gravity Drainage (SAGD) after Cyclic Steam Stimulation (CSS) in a Viscous Oil Reservoir (Unpublished master's thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/32260 http://hdl.handle.net/1880/107038 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
98
Embed
Optimum Switch Timing for Steam-Assisted Gravity Drainage ...
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
2018-06-29
Optimum Switch Timing for Steam-Assisted Gravity
Drainage (SAGD) after Cyclic Steam Stimulation
(CSS) in a Viscous Oil Reservoir
Liu, Siyuan
Liu, S. (2018). Optimum Switch Timing for Steam-Assisted Gravity Drainage (SAGD) after Cyclic
Steam Stimulation (CSS) in a Viscous Oil Reservoir (Unpublished master's thesis). University of
Calgary, Calgary, AB. doi:10.11575/PRISM/32260
http://hdl.handle.net/1880/107038
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
Optimum Switch Timing for Steam-Assisted Gravity Drainage (SAGD) after Cyclic Steam
Stimulation (CSS) in a Viscous Oil Reservoir
By
Siyuan Liu
A THESIS
SUBMITTED TO THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
GRADUATE PROGRAM IN CHEMICAL AND PETROLEUM ENGINEERING
Steam-Assisted Gravity Drainage (SAGD) is often adopted as a follow-up process of Cyclic
Steam Stimulation (CSS) for heavy oil exploitations. Conventionally, CSS is first applied, and
the operation strategy turns to SAGD after the CSS has reached its economic limit. However,
considering the entire well life production, this timing scheme may not be the best. This study
analyzes the optimum timing for the switch from CSS to SAGD. A numerical simulation method
is employed to optimize the timing for CSS to SAGD and several levels of oil viscosity,
permeability, formation thickness, and well spacing are analyzed respectively. A net present
value (NPV) is set as the objective function to evaluate the overall performance and its
sensitivity to the switch point. The optimized operating conditions (temperature, pressure, and
cycle lengths) corresponding to the highest NPV are also generated. In short, this study helps to
maximize the profit of heavy oil development by determining the optimum switch timing from
CSS to SAGD processes.
iii
Acknowledgements
First of all, I would thank my supervisor, Dr. Zhangxing (John) Chen with my sincere gratitude
for supporting my thesis study. I appreciate the valuable opportunity he gave me in this research
group during the past four years.
Second, I give my utmost appreciation to Dr. Xinfeng Jia, who gave me a substantial guidance
on the idea of this thesis. I appreciate his precious time and thought of the assistance he provided
throughout my whole thesis study process.
Third, I have to thank my mother and father, who have been giving me a strong love and mental
support all the time during the past four years. Words can’t describe their great love and
contribution to me during my thesis study.
Fourth, I’m thankful to all the people and friends I met in University of Calgary. I will remember
the pleasant time we spent together and the inspirational moments we shared during the hard
times that passed by. This journey will certainly be remembered throughout my remaining
lifetime.
iv
Table of Contents
Abstract ............................................................................................................................... iiAcknowledgements ............................................................................................................ iiiTable of Contents ............................................................................................................... ivList of Tables ..................................................................................................................... viList of Symbols, Abbreviations and Nomenclature .............................................................x
CHAPTER ONE: INTRODUCTION ..................................................................................11.1 Thermal Methods Introduction ..................................................................................11.2 Problem Statement .....................................................................................................21.3 Thesis Objectives .......................................................................................................21.4 Organization of Thesis ...............................................................................................3
CHAPTER TWO: LITERATURE REVIEW ......................................................................42.1 CSS Literature Review ..............................................................................................4
2.1.1 Overall Introduction ..........................................................................................42.1.2 Production Mechanisms ....................................................................................4
2.2 SAGD Literature Review ...........................................................................................62.2.1 Flow rate control ................................................................................................82.2.2 Influential factors of SAGD performance .........................................................92.2.3 Steam chamber control ....................................................................................10
2.3 Other Techniques .....................................................................................................102.3.1 ES-SAGD process ...........................................................................................102.3.2 VAPEX process ...............................................................................................11
2.4 Reservoir Software Simulators ................................................................................122.5 Information of Liaohe Oilfield’s heavy oil operations ............................................13
CHAPTER THREE: CSS TO SAGD PROCESSES .........................................................153.1 Simulation Model ....................................................................................................153.2 Base Case Study .......................................................................................................173.3 Sensitivity Analysis .................................................................................................24
3.3.1 Viscosity studies ..............................................................................................253.3.2 Permeability studies .........................................................................................293.3.3 Formation Pay Zone Thickness .......................................................................343.3.4 Well Spacing Study .........................................................................................40
CHAPTER FOUR: SAGD TO CSS PROCESSES ...........................................................454.1 Introduction and simulation model ..........................................................................454.2 Base Case Study .......................................................................................................454.3 Results and Discussion ............................................................................................50
4.3.1 Effect of Viscosity ...........................................................................................504.3.2 Effect of Permeability ......................................................................................564.3.3 Effect of Well Spacing ....................................................................................644.3.4 Effect of Payzone Thickness ...........................................................................68
CHAPTER FIVE: CASE STUDY .....................................................................................735.1 Introduction ..............................................................................................................735.2 Field Case Simulation Process .................................................................................755.3 CSS to SAGD results ...............................................................................................775.4 SAGD to CSS results ...............................................................................................795.5 Summary ..................................................................................................................81
CHAPTER SIX: CONCLUSIONS AND RECOMMENDATIONS ................................826.1 Conclusions ..............................................................................................................826.2 Recommendations ....................................................................................................83
Table 3-1 Base Model Parameters ..........................................................................................15
Table 3-2 Parameters for sensitivity analysis .........................................................................25
Table 3-3 Best results for medium and heavy viscosity oil scenario ......................................28
Table 3-4 Best time schemes for six cases for 500mD, 1000mD and 2000mD .....................32
Table 3-5 Results best time scheme cases of 45ft, 60ft and 75ft pay zone thickness scenario ............................................................................................................................38
Table 3-6 Results best time scheme cases of 108ft, 198ft and 288ft pay zone thickness scenario ............................................................................................................................42
Table 4-1 Parameters for Sensitivity Analysis ........................................................................50
Table 4-2 Results for six cases (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) ............................................53
Table 4-3 Results for six cases for 500mD and 2000mD (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) ...59
Table 4-4 Simulation Results of 60ft and 75ft thickness level (upper 60ft; lower 75ft) ........69
Table 5-1 Data of Liaohe field model .....................................................................................74
Table 5-2 NPV calculation data results for each CSS cycle ...................................................77
Table 5-3 NPV table for different timing schemes from CSS to SAGD ................................78
Table 5-4 NPV table for different timing schemes from SAGD to CSS ................................79
Figure 3-9 Chamber chart CSS+SAGD (a) Day 4 of CSS injection period (b) First day of production period, (c) Last day of CSS period, (d) Day 2 of SAGD, (e) Day 190, (f) last day of SAGD period. .......................................................................................................24
Figure 3-10 Two viscosity profiles (Green: heavy, Red: medium) .........................................25
Figure 3-11 Cumulative oil and water production of case 3-yr CSS + 2-yr SAGD (a,b, c, d) .......27
Figure 3-12 Bar graph of NPVs for six cases (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) ......................27
Figure 3-13 Chamber Development at the end of 5 year ........................................................29
Figure 3-14 Cumulative oil and water production for 500mD and 2000mD case ..................30
Figure 3-15 Bar graph and total NPV vs CSS years for 3 permeability scenarios ................32
Figure 3-16 Case comparison for best NPV and CSS years (Permeability) ...........................34
viii
Figure 3-17 Chamber development at the end of the simulation period of the best NPV case ...................................................................................................................................36
Figure 3-18 Bar graph and total NPV vs. CSS years three thickness scenarios (a to f) .........37
Figure 3-19 Case comparison for best NPV and CSS years (Thickness) ...............................39
Figure 3-20 Chamber development at the end of the simulation period (thickness) ...............39
Figure 3-21 Cumulative oil and water production of 198ft well pair spacing (a, b) ...............41
Figure 3-22 NPVs for three well pair spacing levels (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) ..........41
Figure 3-23 Chamber development at the end of the simulation period of the best NPV case ...................................................................................................................................43
Figure 4-1 Pressure control of the process SAGD to CSS (3yr SAGD 2yr CSS) ...................46
Figure 4-2 Typical Oil production curve .................................................................................47
Figure 4-3 Typical water production curve. ............................................................................47
Figure 4-4 Chamber Chart of SAGD to CSS process (SAGD 3yrs + CSS 2yrs) (a to f) ....49
Figure 4-5 Viscosity Effect on Cum. oil and water production (3-yr SAGD + 2-yr CSS) ......51
Figure 4-6 bar graph of NPVs for six cases (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) .........................52
Figure 4-7 Chamber development at the end of 5-year graph .................................................56
Figure 4-8 Cumulative oil and water production of 500mD and 2000mD ..............................57
Figure 4-9 Bar graph of NPVs for six cases (0+5; 1+4; 2+3; 3+2; 4+1; 5+0) .......................59
Figure 4-10 Chamber development at the end of simulation period of 500mD and 2000mD ............................................................................................................................63
Figure 4-11 Cumulative oil and water productions for well spacing of 198ft and 288ft ........65
Figure 4-12 Spacing of 198ft and 288ft spacing NPV simulation results (1 to 4) ..................66
Figure 4-13 Steam chamber phase progression for 198 ft and 288 ft .....................................68
Figure 4-14 Bar graphs and total NPV with CSS years of 60ft and 75ft thickness level. .....70
Figure 4-15 Steam chamber growth of thickness 75ft at the end of 5 years (a to f) ...............71
Figure 5-1 Case Model .............................................................................................................73
ix
Figure 5-2 Field Model History Match ...................................................................................75
Figure 5-3 Total NPV vs. CSS timing and Bar values of CSS and SAGD process ...............78
Figure 5-4 Total NPV vs. SAGD timing and bar values of SAGD to process ........................80
x
List of Symbols, Abbreviations and Nomenclature
Symbol Definition q Oil flow rate ϕ Porosity So Oil Saturation
k Permeability g Gravity acceleration α Thermal diffusivity of rock and fluid h Oil Drainage height m Exponent parameter νo Oil kinematic viscosity NPV Net present value N Total number of periods t Timing of cash flow i Discount rate Rt Net cash flow at month t
xi
1
Chapter One: Introduction
1.1 Thermal Methods Introduction
The twentieth century marks the latest development of the oil and gas industry. Newest oil plays
as well as the most innovative technologies come out gradually within this era. Heavy oil is one
of the most important resources among the non-conventional resources, which comes mainly
from Canada, Venezuela, and United States. However, its recovery process has been challenging
due to its low mobility properties. As the substances of heavy oil are heavy hydrocarbon
molecules, the molecular interaction between each other is complicated, making its viscosity
higher and causing its flow harder than the conventional light oil.
At present, thermal production techniques that are widely used are Steamflooding, Cyclic
producible reserves as well as steam injectivity, and decrease the number of wells required for
field development (Joshi, 1991). However, few pilot test in early 2000s had success in horizontal
well application as operating costs for generating steam still remained high due to a greater heat
loss when steam injection is schemed to a horizontal well application. Creating fractures using
hydraulic fracturing allows a more efficient placement of injected steam, heating up a larger
volume of reservoir and reducing residual oil saturation. This combination is usually considered
for low-permeability heavy oil reservoirs like California diatomite or Athabasca oil sands.
2.2 SAGD Literature Review
SAGD (steam-assisted gravity drainage) is a thermal recovery process which is applied in a large
scale in recent years, invented by Dr. Roger Butler around 1969. It consists of two horizontal
wells which locate in one vertical plane theoretically. Normally the horizontal producer locates
several meters below the injector, and two wells form a heavy oil recovery unit in a specific
reservoir region. A steam chamber is formed after steam is injected to the upper injector. The
recovery of SAGD is as high as 70%.
7
Figure 2-3 SAGD process illustration (Exxon Mobil)
Preheating is an essential part of the SAGD process. Before steam injection, the total area
should be preheated for several months by circulating the steam to the well pattern areas.
The gravity force is the main driving force of the SAGD method. After oil’s viscosity is reduced
and becomes movable, it will flow along the surface of the chamber interface to the producer
well with the steam condensate in the bottom of the chamber (Figure 2-4)
Figure 2-4 Gravity Drainage Theory (Butler, 1991)
Both the horizontal wells locate at the bottom of an oil pay zone. The steam is injected
from the horizontal injector, and starts to rise. The oil phase will receive the conduction heat
8
from the steam and become movable, and the steam becomes condense water as well. Therefore,
the heavy oil and water condensate flow down towards the bottom driven by the gravity force.
The interface between the steam and the cold immobile oil forms and the shape of the steam
zone is like a chamber. After the oil is produced, the steam will occupy the pore space which is
originally taken by oil, and, therefore, the chamber grows gradually, upwards and sideways.
During the process, the chamber pressure almost remains the same (Butler, 1991).
2.2.1 Flow rate control
An oil drainage rate is often described in the following equation (Butler, 1991):
𝑞 = #$Δ%&'()*+,-
(2.1)
which is parallel to the vapor-liquid interface, where q is the oil flow rate to the production well,
φ is the porosity, ΔSo is the average oil saturation change, k is the permeability, g is the gravity
acceleration, α is the thermal diffusivity of rock and fluids, h is the oil drainage height, m is the
parameter, and νo is the kinematic viscosity of oil.
Compared to the typical steamflooding, the SAGD process has a potential of reducing the
steam override (Butler, 1991) as wells are placed in the vertical plane and the gravity force
caused by the density difference of steam and the oil and water condensate phases is totally
beneficial to make the condensate flow to the producers. The viscous fingering effect is reduced
to a minimum level as well.
9
2.2.2 Influential factors of SAGD performance
As the SAGD process is an enthalpy conduction process from a steam chamber to its adjacent
outer cold oil reservoir, the conduction efficiency is the most essential part that needs to be
guaranteed. Therefore, although SAGD is well employed in several parts of the world’s heavy oil
fields, there are also limitations on this process.
According to Edmunds and Chhina (2001), the pay zone thickness plays an essential role
during the SAGD process. The performance will decrease drastically if the reservoir thickness is
less than 15 meters, as the steam chamber may reach the overburden rock during the growing
process, and the enthalpy will disperse to the cap rock layers.
In the geological aspect, the reservoir heterogeneity plays a large role. The existence of
lean zones can heavily influence the shape and the growing process of a steam chamber,
therefore reducing the heat conduction efficiency to the cold oil area. According to Baker et al.
(2008), water zones at the bottom of a reservoir can turn to thief zones, which absorb large
amounts of heat from the steam chamber, increasing the heat loss and limiting the chamber
drainage height. Shale barriers can also become a significant factor that influences the heat
transfer performance. The location of a barrier is essential and the heat conduction is impeded if
it locates in the middle of the horizontal injector and producer. According to Shin and Choi’s
study (2009), a shale barrier can reduce the SAGD performance to the greatest level when it
locates between the injector and the producer. However, it can reach a minimal level when it
locates above an injector, with a vertical distance ranging from 5 to 25 meters.
The oil viscosity is a considerable factor as well. According to Li et al. (2008) the oil
viscosity tends to increase as the reservoir depth increases. However, whether it will exert a
positive or negative effect is not ascertained. Gates et al. believed that SAGD performance
10
deteriorates as the oil viscosity varies vertically (2008) but Chen and Ito (2012) held that by
carrying out a numerical analysis the viscosity variation in the vertical orientation is not obvious.
2.2.3 Steam chamber control
The steam trap control process is an important aspect for operators to consider. Usually there is a
temperature difference between the steam areas around the injector and the producer in a SAGD
pair which is surrounded by the oil and water condensate to be produced. The goal to design this
is to prevent steam from being produced through the bottom producer, so a liquid pool is formed
and submerges the producer. The temperature difference is named subcool, and normally it is
kept around 10 to 20℃to maintain the production (Edmunds et al, 1991). The subcool is
controlled by monitoring the pool’s liquid level (Gates and Chakrabarty, 2008). According to Ito
and Suzuki who carried out a simulation model study in Hangingstone reservoir, Canada (1999),
the Cumulative Steam-to-Oil (CSOR) is minimally reached by controlling the subcool to 30 to
40℃ which is a rare exceptance. As a steam chamber zone is several hundred meters
horizontally in the SAGD process, the subcool should also be monitored along the whole length
of the horizontal well pair (Gates and Leskiw, 2008). Furthermore, a constraint of a producer’s
maximum steam rate should be applied to maintain the steam trap during a SAGD reservoir
simulation process (Gates and Chakrabarty, 2008).
2.3 Other Techniques
2.3.1 ES-SAGD process
ES-SAGD (Expanding-Solvent Steam Assisted Gravity Drainage) is an enhancement of the
traditional SAGD with a small amount of hydrocarbon additive injected along with the steam
11
into an upper injector well (Nasr et al., 2003). The solvent is dissolved into the bitumen at the
chamber edge, reducing the bitumen’s viscosity and making the oil mobile and flow together
with it. It is an enhancement of SAGD and has been successfully field-tested to achieve a better
performance than the SAGD process. Studies show that oil production rates have been improved
as well as OSR (Oil-Steam Ratio) and the water consumption and energy required have been
decreased.
2.3.2 VAPEX process
VAPEX (Vapor Extraction) is also a solvent based process, with a similar well configuration and
solvent chamber with ES-SAGD. The solvents are carefully chosen for this process (e.g., ethane,
propane, butane, etc.). Diffusion plays a vital role in VAPEX while the solvent dissolves into the
bitumen phase and it will process in a slow motion (Das and Butler, 1996). After oil becomes
mobile, it flows along the vapor-liquid interface to the horizontal production well. Butler and
Mokrys (1989) assumed that flow in a mobile oil zone is a laminar flow and the flow streamlines
are parallel to the vapor-liquid interface. The VAPEX is suggested to be a more economical
process especially for thin heavy oil reservoirs as the absence of the heat loss to the overburden
and underburden layers (Butler and Mokrys, 1991) but experiments are still to be implemented
as the diffusion process is too slow, making the production rate lower than that from the normal
SAGD. Also, the cost of a solvent is also a problem to be solved.
12
Figure 2-5 Mechanism of VAPEX process (Butler and Mokrys, 1991)
2.4 Reservoir Software Simulators
Reservoir simulators are the basic computer tools utilizing the high performance of visualization
and computations of CPU cores. They are used in building reservoir models grids, property
iteration calculations, and conveying the overall recovery performance. For CMG’s STARS
simulator, a thermal model is used for its calculations.
Based on the mass conservation law in a reservoir finite unit and diffusivity equations,
basic solutions of oil and water flow rates is obtained in multiphase circumstances in a reservoir.
Different parts of grid blocks contain the information of different phase’s properties, such as
compressibility of rock, permeability, initial water and oil saturation and viscosity. Phase behavior
diagrams are also contained. Numerical techniques such as finite difference formulations are
applied. IMPES (Implicit Pressure-Explicit Saturation) or Fully implicit calculation methods are
options that is adopted during the simulation process.
13
Figure 2-6 Workflow of CMOST module simulation (CMG Manuals, 2016)
CMOST is a module of CMG that is developed to carry out the sensitivity analysis,
optimization, uncertainty analysis, and history match process of reservoir simulations. Using a
base model that is preset, the experimental analysis theory and Monte Carlo simulations are used
to generate different values of several specific reservoir parameters within the preset range,
carrying out simulation using the generated parameter values, analyzing the results and using them
for further parameter value generations (Figure 2-6, CMG Manuals, 2016). Each cycle of the flow
can generate one simulation, or experiment, and every experiment is linked to an objective
function, such as cumulative oil production, an oil rate, or a net present value, which is defined in
the software workflow. After a certain number of experiments are carried out, the best value of an
objective function is adopted as the optimization final result.
2.5 Information of Liaohe Oilfield’s heavy oil operations
Liaohe Oilfield is a field where heavy oil resources consist of a significant portion in China’s main
oil reservoir plays with an overall estimation reserves of 180 million tons. Compared to heavy oil
14
reservoirs in Canada, Liaohe’s oil reservoirs are deeper overall. About 25% of the reservoirs is
within the depth of 600 to 900 meters, 45% within the depth of 900 to 1300 meters, and another
24% portion within the depth of 1300 to 1700 meters. Their heterogeneity is also severe, with a
plenty of reservoirs with bottom water, side water or gas cap, with a large number of faults
contained.
The CSS method started in Liaohe in the 1980s with most wells of vertical configurations.
The SAGD method started to be applied in the late 1990s with dual horizontal well trajectories or
vertical and horizontal well combinations to utilize the previously drilled vertical wells. For some
reservoir cases a recovery factor of more than 30% is achieved after the SAGD approach is
adopted.
Normally CSS is applied with cycles up to twenty, to the circumstance of almost no oil
production, and then shut down operations or other conversion methods are used for following
recovery, which takes a large amount of time, with little cash flow for a single well. Therefore,
we presume there is an optimum timing to concede the CSS process. The switch time from CSS
to SAGD should be optimized to reach a maximum performance of the overall recovery in field
operations. The following chapters will discuss this in more details.
15
Chapter Three: CSS to SAGD Processes
3.1 Simulation Model
The base model is developed to simulate half of the CSS or SAGD process in a typical Canadian
heavy oil reservoir. The basic parameters are listed in Table 3-1.
Table 3-1 Base Model Parameters
Half Model Cross-section area, ft2 54*45 Half Model Length, ft 54 Porosity 0.3 Permeability, k (Darcy) 1000 Reservoir temperature, F 75 Oil zone compressibility, Ct,o, 1/psi 5*10e-4 reservoir pressure, psi 125 Dead heavy oil viscosity @ 75F (cp) 35745 Dead heavy oil density, lbm/ft3 60 Oil saturation, So (vol.%) 0.6 Oil average Thickness, ft 45 Depth, ft 1500
Reservoir model. This homogeneous model is built by the Builder module of the
reservoir simulation software sets CMG, which conprises 270 grid blocks, 18 grid blocks in the
X direction, 1 block in the Y direction, and 15 blocks in the Z direction. Therefore, it is regarded
as a one slice reservoir. With each block’s size of 3×3×3 ft, this model has a dimension of
54ft×3ft×45 ft. So the well spacing is considered as 108 ft in this base case. The initial oil
saturation is set as 0.6 and permeability is set as 1000 mD. The depth of this model is set as 1500
ft.
Wells. All wells in this model are set along the J direction, vertical to the IK plane. For
16
the perforation block, the injector and producer of the CSS well pairs are both located in grid (1,
1, 8) which is the middle block in the K direction. For the SAGD horizontal well pattern, the
injector well locates in (1, 1, 8) and the producer well locates in (1, 1, 15).
Figure 3-1 Reservoir simulation basic model
Viscosity. Figure 3-2 shows the viscosity temperature profile of the two different oil models. The
thermal model from the CMG heavy oil base case is used for viscosity information. A medium
Oil, whose viscosity at reservoir temperature (75F) is 35000 cp and a heavy oil, whose viscosity
at reservoir temperature is 180000 cP, are the two oil models that are used in this study.
17
Figure 3-2 Model Viscosity-Temperature Curves
Relative permeability. The relative permeability’s property of the reservoir fluid is
shown in Figure 3-3 (Detailed data in Appendix 2), where the gas phase is not considered in this
model.
Figure 3-3 Relative permeability curves (Data in Appendix 2)
3.2 Base Case Study
Timing scheme. To determine the optimum timing for starting SAGD after a CSS process, this
study designs a new scheme. First, the net present value (NPV) is set as the objective function of
y=3E+13x-4.923R²=0.98303
y=7E+15x-5.742R²=0.98777
0.10
1.00
10.00
100.00
1000.00
10000.00
100000.00
1000000.00
0 100 200 300 400 500 600 700
Viscosity
(cP)
Temperature(F)
Medium
ViscHeavy
18
the optimization process. Second, a total period of 5 years of the CSS-after-SAGD process is
stimulated. Third, six cases or six combinations are considered:
(1) 5-year CSS
(2) 4-year CSS and 1-year SAGD
(3) 3-year CSS and 2-year SAGD
(4) 2-year CSS and 3-year SAGD
(5) 1-year CSS and 4-year SAGD
(6) 5-year SAGD
After the simulation of these six cases, the NPVs for each of the case are compared, and the
optimum timing is determined.
Operation scheme. Pressure control is one of the most essential operation conditions.
shows the pressure change for the case of two years CSS and then switched to SAGD. It is noticed
that in the injection period of the CSS process, the pressure starts up high at about 950 psi, and
drops off drastically during the soaking period and production period as the oil and water starts to
come up to the surface. Two cycles for the CSS have been implemented and the pressure cycles
also reflects this. The pressure for the last three years remains constant as during the SAGD period.
The process of chamber growing with the steam injection and oil production from the upper
injector and lower producer, respectively, is consistent throughout three years.
19
a) Producer well block (1, 1, 15) b) Injector well block (1, 1, 8)
Figure 3-4 Pressure control of the 5-year simulation process
NPV. NPV is set to be the economic indicator of the entire process. Regarding to a certain
specific month, it is set as: NPV= Oil per volume price*(CSS Total Oil Production + SAGD Total
Oil Production) - Steam per volume cost*(CSS Total Steam Used + SAGD Total Steam Used).
For a total simulation period of 5 years, The equation is used as below:
𝑁𝑃𝑉 𝑖, 𝑁 = 34567 4
89:; (3-1)
𝑁 – Total number of periods (The number of total processed months at a specific study point).
𝑡 – Timing of the cash flow (The month’s number at the study point).
𝑖 – Discount rate (a 0.007 monthly discount rate is used).
𝑅9 – Net cash flow (oil revenue – water cost) at month t.
It is worthy of noting that $60 per barrel of an oil revenue and $5 per barrel of a water cost are
used. The operation and other operation costs are neglected in this model.
Constraints. Constraints are set for the CSS and SAGD processes: (1) CSS: bottom hole
pressure 1004.5 psi for the maximum and a surface water rate of 5000 bbl/day for the maximum
20
for the CSS injector; bottom hole pressure 200psi for the minimum for the CSS producer. (2)
SAGD: producer: bottom hole pressure 225.4psi for the maximum and a surface water rate of 2
bbl/day for the maximum for the SAGD injector; bottom hole pressure 215.4 psi for the minimum
and a surface liquid rate of 5 bbl/day for the maximum for the SAGD producer.
Performance. The production curves of CSS and SAGD is generated. For the CSS period,
there are two peaks in the oil production rate curve as the blue line in Figure 3-5 (a) shows. The
highest rate is about 12 bbl/d. However, the overall producing time is very short. The cumulative
oil production for the total two CSS cycles is around 38bbl.
After the CSS well pairs shut in and the whole process convert to the SAGD process, the
oil production continues and from the blue line in Figure 3-5 (b), the oil production rate becomes
more consistent and prolonged compared to the previous CSS period. It increases gradually and
the highest oil production rate reaches 0.26 bbl/d at approximate 1 year after the start of the SAGD
process. It decreases in a relatively smooth fashion after the peak point.
a) CSS b) SAGD
Figure 3-5 Oil cumulative production and rate (a: CSS; b: SAGD)
The cumulative oil production for the SAGD period is 110 bbl, which is much higher than
21
the CSS period and gives a better recovery performance for the residual oil left in this model
(Figure 3-6). The water production also expresses a similar shape. For the CSS process the
maximum water production rate for the total two cycle is about 980 bbl/d according to the
simulation results and also continues in a very short production period similar to the oil production
curves. For the SAGD period it reaches about 1.4 bbl/d and comes to 1 bbl/d in the final period.
So the cumulative water production is growing consistently in the 3-year process.
a) CSS b) SAGD
Figure 3-6 Oil and water production of SAGD (a: CSS; b: SAGD)
The CMOST module is a history match and experimental parameter sensitivity analysis
tool in the CMG simulation software. In this study it is utilized to carry out the NPV optimization
process. By running it, the operation of running STARS for a single process is repeated, meanwhile
manipulating the parameters that control the thermal process, such as injection steam pressure,
steam temperature, and a producer well steam rate. The total number of experiments is set in
CMOST and after all the experimental simulation cases are run, the case with the highest NPV is
generated.
Normally up to 200 total experiments is carried out for a single process optimization, if a
22
stable trend for the maximized NPV solution is reached. From Figure 3-7 the optimal solution
with the red dot is the solution with the highest NPV among all the experiments that have been
carried out. The two-year CSS in this base case has a highest NPV for about $300 shown in Figure
7.1 and for the SAGD case the highest NPV result is about $1500.
Figure 3-7 CMOST Optimization graph (CSS is Upper, SAGD is lower)
Optimization process. The model runs for 5 years in six schemes: 0-year CSS 5-years
Figure 4-15 Steam chamber growth of thickness 75ft at the end of 5 years (a to f)
4.4 Chapter Summary
This chapter illustrates the outcomes that SAGD simulation is implemented first and then
converted to the CSS method for the entire 5 years simulation period. Overall several points is
generated from the simulation results:
According to all various scenarios, it is better to implement SAGD at least 3 years to produce the
greatest NPV for the simulation model for all the six time schemes analyzed. Effects of analyzed
parameters exert in the NPV results:
1. Permeability: Increasing of permeability can increase the NPV value in regards to the same
72
time scheme (except for 5 years CSS). Compared to the CSS to SAGD process in the
previous chapter, theSAGD to CSS process delivers a higher NPV for the highest time
schemes.Viscosity: For each time scheme in the SAGD to CSS process, heavy oil’s NPV
result is lower than the counterpart of the medium oil NPV result, which indicates that a
viscosity increase leads to a decrease in the overall NPV performance. Compared to the
CSS to SAGD process, SAGD to CSS takes advantage in the medium oil scenario for the
maximum NPV time scheme, while CSS to SAGD reaches a better NPV for the heavy oil
scenario.
2. Payzone thickness: for the three different scenarios of thickness analyzed for the total NPV
value, the SAGD to CSS process takes advantage in the 45 ft and 60 ft thickness levels by
a relatively big margin (4yr SAGD 1yr CSS vs. 1yr CSS 4yr SAGD for 45ft thickness, 4yr
SAGD 1yr CSS vs. 1yr CSS 4yr SAGD for 60ft thickness), while CSS to SAGD takes
advantage in the 75 ft level(2yr CSS 3yr SAGD vs. 4yr SAGD 1yr CSS).
3. Well spacing: CSS takes more NPV percentage as the well pair spacing increases from 108
ft (the base case) to 288 ft. At the 288 ft level, only CSS can generate a positive NPV value,
which is similar to the previous chapter’s results. Also, taking the well grids into account,
as the 198 ft level’s total NPV is greater than 108ft’s for most time schemes, the 5 years’
total NPV generated per reservoir volume will decrease as the well pair spacing increases.
73
Chapter Five: Case Study
5.1 Introduction
This chapter focuses on simulation of a field model of a heavy oil reservoir, which uses the similar
approach of CSS to SAGD and SAGD to CSS processes with all six timing schemes adopted in
the previous chapters’ analysis, and a rough comparison is made between the results of this field
case model and previous homogeneous models. The NPV value of each single thermal process as
well as the total NPV value for all the timing schemes are generated as well.
The previous two chapters illustrate the best optimization timing scheme under different
reservoir parameters. This chapter investigates the validity of the optimization scheme in a real
heavy oil reservoir. The geological model is from Liaohe Oilfield in China, which is of a different
kind from the previous chapters’ models as it is a heterogeneous model.
Figure 5-1 Case Model
The heavy oil reservoir model has 12×14×8 grids. It is a deep heavy oil reservoir with an
average depth of 1340 meters. The oil saturation of every blocks of this model ranges from 0.47
74
to 0.75, the average pay zone thickness is 15 meters, and permeability ranges from 800 mD to
1000 mD horizontally, 70 mD to 300 mD vertically, and 80 mD to 100 mD for the most middle
part. The viscosity of oil in this reservoir is around 5000 mPa.s at reservoir situations. Horizontal
well pair spacing is around 70 meters.
The shape of the reservoir is declining downwards, and two horizontal wells are located in
this reservoir, with the upper well used as both the CSS well and SAGD injectors, and the lower
well as SAGD horizontal producers. Due to the decline, the two well trajectories and perforations
are not fully overlapped horizontally in the JK plane. More information of the reservoir model is
listed below (Table 5-1).
Table 5-1 Data of Liaohe field model
Average IJ Cross-section area (m) 100*70 Length (m) 100 Porosity (%) 0.178-0.247 Horizontal Permeability, k (mD) 800-1000 Vertical Permeability, k (mD) 80-300 Reservoir temperature, T (C) 50C Surface pressure, Psc (kPa) 1.01325*102 reservoir pressure (kPa) 1800 Dead oil viscosity (mPaS, at 30°C) 9926 Dead oil density (g/cm3) 0.987 Oil Saturation, So (vol.%) 0.52 Oil Thickness(average, m) 15 Depth (m) 1320-1356
The history match is also carried out between the simulation results of the original placed injectors
and producers of this model with the original production history for the previous 16 years (year
1993-2009). From the results of the field cumulative oil production (Figure 5-2), it can be seen
that the simulation result for most of the period gives a good match of the production history, in
75
terms of its trend and the cumulative oil number, which proves its validity of this field model’s
petrophysical properties for further studies.
Figure 5-2 Field Model History Match
Overall, the reservoir properties are covered in the case studies in Chapters 3 and 4:
medium oil viscosity, permeability of 1000 mD, well spacing of 198 ft, and payzone thickness of
45ft. Therefore, conclusions from the previous chapters is referred to determine the operating
scheme.
5.2 Field Case Simulation Process
The goal of doing this Liaohe field model simulation is to show that the screening method for the
best timing scheme to convert from CSS to SAGD in this thesis delivers a better economic value
than the scheme used normally in field operations.
During heavy oil operations, CSS operations are frequently used as initial stage thermal
methods prior to other methods as it is relatively more applicable. When the economic limit is
reached, CSS operations is terminated and a follow-up recovery method is applied, such as SAGD.
76
In this thesis, the CSS termination point is altered in an overall view of the total five-year recovery
period to increase the economic outcome to the maximum amount.
In order to set the termination criteria of CSS operation, the economic profit is analyzed.
For this analysis, each month is used as a timing unit and each CSS cycle is regarded as one
evaluation timing period to calculate the Net Present Value (NPV). If one evaluation timing
period’s NPV turns from a previous positive value to a negative value, this cycle’s timing implies
that the CSS operation should be terminated and it is the timing to switch to the SAGD process.
NPV is analyzed in the equation below.
𝑁𝑃𝑉 𝑖, 𝑁 = 34567 4
89:; (5-1)
𝑁 – Total number of periods (in this case it is the total months in the five year simulation period)
𝑡 – Timing of the cash flow (In this case it is the month’s number)
𝑖 – Discount rate (a monthly discount rate is used, converted from 0.1 yearly discount rate which
used in CMOST is 0.007)
𝑅9 – Net cash flow (oil revenue – water cost) at month t (60$/bbl adopted for the oil price, 5$/bbl
for the water price)
It is assumed that CSS is the method that field operations are going to carry out for the total
five years of the studying period, and, therefore, the CMOST module is utilized for the five-year
CSS timing scheme to find the best case available for CSS operations. The cycle in which NPV is
generated indicates within which cycle NPV starts to turn from a positive value to a negative one.
When a negative value emerges, the cycle is considered to be termination timing.
In this reservoir model, after simulation using CMOST for the 5-year CSS and 0-year
SAGD scheme, 5 cycles are generated. All the present value is carried out for each cycle (Table
77
5-2), and cycle 4 is the one in which the NPV value starts to turn negative. So, based on the
previous assumption, Cycle 4 should not be executed in field operations and CSS should be
terminated at the third year. Therefore, 3-year CSS and 2-year SAGD is the timing scheme that
should be adopted in field operations.
Table 5-2 NPV calculation data results for each CSS cycle
Figure 5-4 Total NPV vs. SAGD timing and bar values of SAGD to process
Also, the trend line analysis method is utilized in this SAGD to CSS process of the field
model case study to account for the influence of other timing schemes’ total NPV results. An
equation of y = -1908.5x2 + 12160x + 5367.3 (R² = 0.8105, y: total NPV value, x: SAGD years)
is generated. In this case, the maximum NPV value is reached after implementing SAGD for
3.18 years, which means that the conversion point from SAGD to CSS is reached at the second
month of the fourth year within the 5-year simulation period. For the resemblance with the
previous 198ft well spacing homogeneous case of SAGD to CSS (Section 4.3.3 in Chapter 4),
there exists some major difference in the timing scheme of one year SAGD four years CSS, and
the five year SAGD timing scheme, but overall the similarity still remains as the SAGD and CSS
both takes up a certain amount for the other timing schemes.
y=-1908.5x2 +12160x+5367.3R²=0.810530
10000
20000
30000
40000
0 1 2 3 4 5
TotalN
PV($)
SAGDTimingPeriod(yr)
TotalNPVvs.SAGDTiming
0
10000
20000
30000
40000
0 1 2 3 4 5
TotalN
PV($
)
SAGDTimingPeriod(yr)
SAGD CSS
81
5.5 Summary
Though some major discrepancies between each other exist, the overall results show the
rationality and feasibility of the NPV analysis method that was adopted throughout this chapter.
For the CSS-to-SAGD process, using this optimization method can find the best timing scheme
that gives a better switching point from CSS to SAGD to generate more revenues than that
adopted in field operations.
Also, for SAGD-to-CSS, 4-year SAGD and 1-year CSS can generate the largest NPV for all the
timing schemes, with SAGD contributing to most of the NPV.
The conclusions from CSS-to-SAGD with homogeneous models are worthy of referring to for
real heterogeneous cases.
82
Chapter Six: Conclusions and Recommendations
6.1 Conclusions
This thesis implemented the analysis of the CSS to SAGD process in a thin sliced homogeneous
3-D reservoir model, with the impact of four different reservoir parameters, viscosity,
permeability, pay zone thickness and well pair spacing, for a total of five-year recovery
simulation. The SAGD to CSS process is also carried out as a contrast study with these four
reservoir parameters. Finally, a heterogeneous field heavy oil reservoir model from Liaohe
Oilfield is utilized to validate the conclusions acquired from the homogeneous model study. The
following conclusions is made from this thesis:
1. For the CSS to SAGD process, optimizing the 5-year recovery process within the scope of
the whole period, by implementing six different timing schemes, is able to give a better
solution of the switching time than carrying out the CSS process to its economic limit and
then switched to the SAGD recovery process, regarding the highest total NPV of the whole
process reached.
2. For the effect of permeability, viscosity, pay-zone thickness, and well pair spacing, each
exerts a certain effect on the total NPV and total oil recovery, and also the recovery
percentage of each thermal process is changed with different parameters of a reservoir
model taken into simulation.
3. For the SAGD to CSS process, the overall effect of the previous four reservoir parameters
gives a similar impact to the total NPV result. The total NPV value of SAGD to CSS is
higher than CSS to SAGD for most reservoir parameter scenarios.
4. The switch timing from CSS to SAGD is estimated to a higher extent by using the
83
polynomial trend line regression method. Also, the best CSS to SAGD switch time for other
reservoir properties (viscosity, permeability, pay-zone thickness, and well spacing) is
roughly estimated, which is also beneficial for forecasting the switching point for a variety
of different reservoir parameter values.
Overall, this optimization method is a novel way to optimize the recovery strategy of certain
heavy oil reservoirs.
6.2 Recommendations
In future, the following research work is recommended to improve the optimization strategy for
multiple thermal recovery processes:
1. Larger ranges of operating conditions should be considered, for example, a range of steam
temperatures for CSS and SAGD.
2. More thermal recovery processes, such as steamflooding, hot/cold water flooding, and in
situ combustion, is included in next step studies.
3. More sensitivity analysis should be conducted. For example, two oil viscosity profiles
properties are investigated, but in future more scenarios should be considered.
84
References
Edmunds, N.R., Kovalsky, J.A. Gittins, S.D., and Pennacchio, E.D., Review of the Phase A Steam Assisted Gravity Drainage to un Underground Test Facility; SPE 21529, presented at the SPE International Thermal Operation Symposium, Bakersfield, CA, February 7-8, 1991. Green D.W., Willhite G.P. Enhanced oil recovery, SPE textbook series, vol. 6, Henry L. Doherty Memorial Fund of AIME, Society of Petroleum Engineers, Richardson, Texas (1998) Butler, R.M. Thermal recovery of oil and bitumen. United States: N.P., 1991. Web. Edmunds, N., & Chhina, H. (2001, December 1). Economic Optimum Operating Pressure for SAGD Projects in Alberta. Petroleum Society of Canada. doi:10.2118/01-12-DAS Baker, R. O., Fong, C., Li, T., Bowes, C., & Toews, M. (2008, January 1). Practical Considerations of Reservoir Heterogeneities on SAGD Projects. Society of Petroleum Engineers. doi:10.2118/117525-MS Gates, I. D., Chakrabarty, N., Moore, R. G., Mehta, S. A., Zalewski, E., & Pereira, P. (2008, September 1). In Situ Upgrading of Llancanelo Heavy Oil Using In Situ Combustion and a Downhole Catalyst Bed. Petroleum Society of Canada. doi:10.2118/08-09-23 ITO, Y., SUZUKI, S. and YAMADA, H., Effect of Reservoir Parameters on Oil Rates and Steam Oil Ratios in SAGD Projects, presented at the 7th UNITAR International Conference on Heavy Crude and Tar Sands, Beijing, China, 28-31 October 1998 Joshi, S. D. (1991). “Thermal Oil Recovery With Horizontal Wells (includes associated papers 24403 and 24957).” Journal of Petroleum Technology 43(11): 1302‐1304. Nasr, T. N., Beaulieu, G., Golbeck, H., & Heck, G. (2003). Novel Expanding Solvent-SAGD Process" ES-SAGD". Journal of Canadian Petroleum Technology, 42(01). Ito, Y., & Suzuki, S. (1999). Numerical simulation of the SAGD process in the Hangingstone oil sands reservoir. Journal of Canadian Petroleum Technology, 38(09). Butler R.M., Mokrys I. J. (1989). "Solvent Analog Model of Steam-Assisted Gravity Drainage." AOSTRA Journal of Research, 5, 17-32. Das, S. K., & Butler, R. M. (1996). Diffusion coefficients of propane and butane in Peace River bitumen. The Canadian journal of chemical engineering, 74(6), 985-992. CMOST Manual, CMGL Company, 2016
85
Appendix
1. Units used in Tables
CSS years
Timing Schemes
Inj Temp.
Max I. Timing
Max S. Timing
Max P. Timing NPV Inj
Temp Prod. Rate NPV Total
NPV
CSS years years
Inj Temp. Fahrenheit
Max I. Timing days
Max S. Timing days
Max P. Timing days
NPV Dollar ($)
Prod. Rate bbl/day
Total NPV Dollar ($)
2. Value of Relative Permeability Curves Value of the Basic Simulation Basic Model