SAGD Simulation Modelling
using STARS
CMG Reservoir Simulation Seminar
September 20-21, 2011 Suzhou, China
CMG UnionST
Presentation Agenda
What is SAGD
Thermal Recovery Process Overview
Process Mechanisms
Description of Reservoir
Description of Fluid Components
Description of Thermal Properties
Description of Rock-Fluid Properties
Description of Wells ( Flex well presentation)
13-
What is SAGD?
SAGD is Steam Assisted Gravity Drainage
Gravity driven process, NOT a vertical steam flood
Implications:
Injector producer block pressure difference can be zero
Oil rate depends on viscosity reduction, formation permeability,
and drainage slope
What is SAGD?
A horizontal well pair is drilled near the bottom of the reservoir
The upper well is used to inject steam continually to form a
growing steam chamber
At the steam-oil interface, steam releases its heat energy to the
adjacent oil zone
As a result, steam is condensed to water and oil is mobilized
The mixture of oil and water condensate is drained and removed by
gravity at the lower horizontal well
As the oil is removed, the steam chamber grows upward and laterally
13-
What is SAGD?
High recovery factor (60 to 80%) and favorable energy
efficiency can be achieved
Both injection and production wells are usually operated at
a constant pressure
The process has been modified into many configurations,
such as single well SAGD & a combination of vertical
injectors and horizontal producers
The process may be operated (with care) in reservoirs that
contain gas cap and bottom water
What is SAGD?
Steam Chamber growth in SAGD
13-
Process Mechanisms
Unique characteristics of water and steam
Water and steam are chemically stable
They are excellent heat transporting agents
h = hw + X Lv
h = specific enthalpy of the wet saturated steam
hw = specific enthalpy of water
X = steam quality fraction
Lv = latent heat of vaporization
hw increases with steam pressure while Lv
decreases with steam pressure
Process Mechanisms
Phase changes of hydrocarbon
Phase behavior of hydrocarbons in thermal processes
is much more complex than that of water
For many heavy oil processes, 2 pseudo-hydrocarbon
components are sufficient: dead oil & solution gas.
Dead oil exists only in liquid phase and solution gas
can exist in both liquid and gas phases
13-
Process Mechanisms
The effect of thermal expansion and compressibility
of liquid
Density of liquid changes with temperature and
pressure in the reservoir
A temperature increase would increase the liquid
volume, while a pressure increase would decrease the
liquid volume
The overall effect of temperature and pressure change
on density is relatively small because these two
factors negate the effect of each other to a certain
extent
Process Mechanisms
The effect of rock compressibility
change due to solid expansion
compression is insignificant because solids are
characterized by their low coefficient of thermal
expansion and coefficient of compressibility values
porosity and permeability properties of
unconsolidated sands can be altered substantially by
the change in temperature and/or pressure
13-
Process Mechanisms
Formation dilation & fracturing
In some steam operations; injection pressures can exceed
the formation dilation onset pressure or
fracturing pressure in order to establish steam injectivity
Steam enters the formation via dilated zones or
narrow high permeability zones(oriented vertically or
horizontally) depending on the stress condition of the
reservoir
Shallow formations will tend to have horizontal fractures
Deep formations tend to have vertical fractures
‘Unconsolidated’ sands are more affected by dilation
Several constitutive & empirical geomechanical
models have been designed to simulate this
fracturing process
Dilation-Compaction
Formation dilation & fracturing
The analytical dilation-recompaction model in STARS
13-
Process Mechanisms
Heat and mass transfer in porous medium
Both solid and fluid phases contribute to the heat
transfer process when the fluids flow through a non-
isothermal porous medium
When fluid rate is slow, the conduction process has a
strong effect on the heat transfer
When fluid rate is high, convection dominates the
heating process and the thermal conductivity can
become less important
Heat capacity is always important
Process Mechanisms: Thermal
Conductivity
Averaging Methods (THCONMIX): Thermal conductivity
determines the flow term K∆T due to diffusion of energy from a region of high
temperature to low temperature.
Simple: Simple volume
weighting of phase thermal
conductivities.
Complex: More complex
mixing of phase thermal
conductivities.
Porosity = 0, the rock value is
use.
0 < PHI <0.01 When porosity is
between zero and 0.01, the
calculation assumes 0.01.
13-
Process Mechanisms: Thermal
Conductivity
Process Mechanisms: Thermal
Conductivity
Simulation Field example differences when using Complex vs.
Simple average thermal conductivity methods.
SAGD Single well simulation
13-
Process Mechanisms: Thermal
Conductivity
3D View of the Shale Bodies
13-
Reservoir Description
Reservoir Description
10
13-
Reservoir Description
Shale bodies cause a reduction in vertical communication, thus causing a
reduction in the steam injection rate, and oil production rate
Long shale bodies are worse than short shale bodies
SAGD Performance
400
Short Shale Bodies Occupying 10% Volume
Short Shale Bodies Occupying 30% Volume
Long Shale Bodies Occupying 30% Volume
300
Oil Rate SC (m3/day)
200
100
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Tim e (Date)
800
Short Shale Bodies Occupying 10% Volume
Short Shale Bodies Occupying 30% Volume
Long Shale Bodies Occupying 30% Volume
Steam Injection Rate SC (m3/day)
600
400
200
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
Tim e (Date)
STARS Fluid Model Description
Components
Defined by specifying molecular weight, density, compressibility,
thermal expansion, and viscosity as well as critical properties
Liquid properties
Calculated by applying mixing rules to pure component density
and viscosity data
Vapor density
Calculated using a simple cubic EOS
Vapor viscosity
Calculated using a separate compositional viscosity model
Component distributions
Determined from component K-values
11
13-
STARS Fluid Systems Modeled with WinProp
Black oil system
Oil and solution gas
Thermal black oil system
Oil and solution gas; water/steam
VAPEX system
Oil and solution gas; light hydrocarbon solvent
Steam-additive system (enhanced drive or ES-
SAGD)
Oil and solution gas; CO2; light or intermediate
hydrocarbons (e.g. Naphtha); water/steam
STARS PVT Data to Be Generated
Component critical properties, molar densities,
compressibilities, thermal expansivities, and
liquid viscosities
Component K-value tables
12
13-
Description of Fluid Components
¾ One of the first design decisions is the number of components,
type and the phases in which each component can be found
Black-Oil System or Thermal Black-Oil System
Components Phase
Aqueous Oleic Gaseous
Water X
Dead Oil X
Solution Gas X X
Steam-Additive System
Components Phase
Aqueous Oleic Gaseous
Water X X
Heavy Oil X
Light Oil X X
Naptha X X
Carbon Dioxide X X X
Description of Fluid Components
Component design concepts
Water components
Water is a standard component
Internal defaults are available for vapor-liquid K values,
density, viscosity and heat capacity
Stone’s model for oil relative permeability curves assumes
the aqueous phase is the wetting phase
CMG STARS allows for other components besides water in
the aqueous phase
13
13-
Description of Fluid Components
Component design concepts
Oil Components - pure and pseudo-components
A pure component consists of a single molecular species
A pseudo-component is a group of molecular species,
lumped together with a fixed distribution
A native oil may be split into pseudo-components
using the different techniques
Description of Fluid Components
Component design concepts
Gas components - condensable and non-
condensable
A condensable gas is a gas component which is soluble in a
liquid phase. Properties in both the liquid and gas phases
must be supplied, along with K values
A non-condensable gas is a gas component for which the
solubility in liquid is small enough to be ignored. In this case,
K values and liquid property data are not required
Typical examples for non-condensable gas are oxygen,
nitrogen and carbon monoxide in combustion process.
Carbon dioxide and methane may be condensable or non-
condensable, depending on which processes are of interest
14
13-
STARS Component-Phase Chart
Components Phase
Aqueous Oleic Gaseous Adsorbed Solid
Water-like Condensable X X X X
Oil-like Condensable X X X X
Non-condensable Gas X X
Solid X
What STARS needs
Fluid Phase Equilibrium
STARS requires phase equilibrium ratios (K-values)
from thermodynamic model (e.g. PR EOS) or
input directly as functions of pressure, temperature
and composition
go yi xi
gw y K = K ow =
K = i i xi i wi
i wi
15
13-
Description of K Values
Three ways to input gas-liquid K values in STARS
1. Enter 5 coefficients for the correlation
ai di
Ki = ( + b i p + c i ) exp ( )
p T − ei
Table 2 in the STARS User Guide contains values for these coefficients
for many components typically used in thermal simulation.
2. Enter a table of Ki versus p and T. This is useful when K values
come from a series of equation-of-state calculations.
Interpolation between pressure points is the ‘a/p’, while
interpolation between temperature points is the ‘exp(d/T)’
3. Enter a table of Ki versus p, T and the composition of one key
component
Description of Densities
¾ Water and oil phase
Liquid densities are obtained by ideal mixing of pure component
densities with phase composition
nc nc
1 wi 1 xi
ρw
= ∑ ρ wi ρo
= ∑ ρ oi
1 1
• Densities ρw and ρo are inverses of phase molar volume
¾ Gas phase
The mole density of the gas phase is calculated
internally from
ρg = p / RTZ
16
13-
Description of Water Viscosity
Water phase - water phase viscosity tends to vary
between 1 cp at low temperature to 0.1 cp at 300
°C
There are 3 ways to input water phase viscosity in
STARS
1. Use internal table with a possible dependence on salt
concentration which can be significant
2. Use the correlation μw = a • exp (b/T), where ‘T’ is in absolute
degrees
3. Enter directly a table of μw versus T
Description of Gas Viscosity
Gas phase - There are three ways to input gas phase viscosity in
STARS
1. Use the correlation μg = 0.0136 + 3.8 • 10-5 • T, where ‘T’ is in degrees C
2. Use the correlation μgi = ai Tbi (T in absolute degrees) for each component.
The phase viscosity is nc
∑ ω i μ gi
μg = i =1
nc
,
∑i =1
ωi
ω i = yi × M i
3. The third method is the same as in option 2, but with an additional correction
to μg for high pressure.
17
13-
Description of Liquid Viscosity
The liquid viscosity may be specified using one of
the following two options:
1. Correlation μoi = ai • exp(bi/T) where T is in absolute
degrees
2. Enter directly a table of μoi versus T
Rock Heat Capacity
Rock heat capacity is one of the most important
thermal properties for describing the thermal
behavior of a reservoir
The heat capacity of the rock is treated on a
volumetric basis because rock is handled by volume
and not by moles
Rock heat content is described by:
Ur = [a + bT] [T – Tr]
where a = constant volume heat capacity and b =
temperature coefficient
A value of a = 35 Btu/ft3-F is used commonly
18
13-
Thermal Conductivity
Thermal conductivity determines the flow term KΔT due to
diffusion of energy from a region of high temperature to low
temperature
The only other way for energy to flow in situ is by convection
In field scale problems, convection usually dominates conduction,
at least in the direction of flow
Conduction can play a significant role in both steam and
combustion process at the laboratory scale, since the length scale
is much smaller than in the field
Overburden Heat Loss
A semi-analytical model is used for heat transfer to or from an
adjacent formation of infinite extent. It assumes a temperature
profile in the base or cap rock as a function of time and distance
from the reservoir interface
( )
T (t, z ) = θ + pz + qz 2 e − z / d
where
d = diffusion length = √λt/2
θ = gridblock temperature
z = distance
λ = thermal diffusivity
This profile satisfies both T(t,0) = θ and T(t,∞) = 0
The only values required for input are heat capacity and thermal
conductivity of the over/under burden rock
19
13-
Description of Wells
The effect of well bore pressure drop and heat transfer between
tubing and annulus fluids can have a strong impact on the
horizontal well thermal processes such as SAGD
It is desirable to have uniform injection and production along the
horizontal well bores to maximize the reservoir conformance
The following well models will be discussed in this course:
Overview of Flex Wellbore
Source/sink well with G/F/HLOSS option
Uniform grid with LAYERXYZ
Non-uniform grid with LAYERXYZ
Work Flow in SAGD Modelling
Flxwellbore
20
13-
Heater Set-up
The best approach is the 4-component heater
HEATR
Constant value can be used on its own or with TMPSET and
UHTR
UHTR
Variable value which diminishes to zero as TMPSET is
approached Energy = UHTR*(TMPSET-T)
Usually set to 1/10 the value of HEATR to allow it to start 10 C
below the set point temperature
TMPSET
Temperature target
AUTOHEATER
ON – Minimum of HEATR and UHTR used initially then UHTR
takes over until TMPSET is reached
OFF – results in HEATR and UHTR being added together
The Geomechanics Option
21
13-
1000 mD vs. 100 mD Kv at the Same Time
Normal vs. Top Water
22
13-
10 MPa – No Gas vs. High Rs
Gas Cap – Low Rs (1.175 m3/m3)
23
13-
Grid Size Aspect Ratio
Aspect ratios different from 1 also cause distortion
of the steam chamber growth and different oil
production profiles
Temperature profiles indicate not too much
difference between the cases with different aspect
ratiosp ( ) y p ( ) y
0 100 0 100
1x1 Meter Case 5x1 Meter Case
0 100 0 100
Grid Size Aspect Ratio
Production profiles show very different behavior
Corrections can be used if an aspect ratio different from 1.0 must be used
With a 2x1 case, the following corrections are required to match the 1x1 case:
TRANSI = 2, TRANLI =2, and directional thermal conductivities (I-direction * 2)
LATERAL GRID CELL ANISOTROPY
1 x 10 x 1 m VS. 2 x 10 x 1 m VS. 5 x 10 x 1 m CELLS
PRODUCER 28% & 55% ERROR
800
600
2x10x1 m CELLS WITH TRANS*2
1x10x1 m CELLS
2x10x1 m CELLS
Cumulative Oil SC (m3)
5x10x1 m CELLS
400
200
0
2007-4 2007-7 2007-10 2008-1 2008-4 2008-7 2008-10 2009-1
Time (Date)
24
13-
Spacing
STARS TEST SAGD STARS TEST SAGD
2M OIL 5M OIL 10M OIL 15M OIL 50M OIL 100M OIL 150M OIL 200M OIL
2M SOR 5M SOR 10 M SOR 15M SOR 50M SOR 100M SOR 150M SOR 200M SOR
800 5.0 1,000 4.0
4.0 800
600 3.0
Steam Oil Ratio Cum SCTR (m3/m3)
Steam Oil Ratio Cum SCTR (m3/m3)
Cumulative Oil(OIL) SC (m3)
Cumulative Oil(OIL) SC (m3)
3.0 600
400 2.0
2.0 400
200 1.0
1.0 200
0 0.0 0 0.0
1992 1994 1996 1998 2000 2002 1992 1994 1996 1998 2000 2002
Time (Date) Time (Date)
Vertical Horizontal
Spacing Spacing
Temperature and Compositinal
Dependence of
Relative Permeability
KRTEMTAB
Diagnostic plots
25
13-
Temperature Dependent End Points
Trends are the same for different oils and studies,
but actual values are different
Recommend using something like the following:
KRTEMTAB SWR SORW SORG
**$ TEMP (Deg C) SWR SORW SORG Table
12.0 0.13 0.39 0.45 values at
260.0 0.25 0.20 0.05 reservoir
temperatur
e
Steam Residual oil Residual oil
temperature to hot to steam
waterflood flood
SAGD Overview
Horizontal injector and horizontal producer
Steam Chamber
Steam rises, gives up heat
Cold Oil Sand
Horizontal Injector
Hot Bitumen Drops Down
(mobilized bitumen and condensed steam)
Horizontal Producer
26
13-
SAGD Overview
Single well SAGD
Cold Oil Sand
Horizontal Injector / Producer
SAGD Overview
Combination of vertical injectors and horizontal producer
Vertical Injector Vertical Injector
Cold Oil Sand Horizontal Producer
27
13-
SAGD Overview
Combination of vertical injectors and horizontal producer
Vertical Injector Vertical Injector
Cold Oil Sand Horizontal Producer
SAGD Overview
Fast SAGD (adding offset horizontal producers CSS)
Steam Chamber
Steam rises, gives up heat
Horizontal Injector
Horizontal Producer Horizontal Producer
CSS CSS
Horizontal Producer
28
13-
SAGD Overview
Fast SAGD (adding offset horizontal producers CSS)
Steam Chamber
Steam rises, gives up heat
Horizontal Injector
Horizontal Producer Horizontal Producer
CSS CSS
Horizontal Producer
Thank You
for your attention
29
13-
