You are on page 1of 39

L5 Recovery

Processes
ENPE 525/624
Primary Recovery Processes
• Pressure Depletion • Water Influx
• Gas Cap Expansion • Side or Bottom Water Influx
• Solution Gas Drive • Water Alternating Gas
• Foamy Oil Production
Secondary Recovery Processes

• Gas Injection • Water Injection


• Immiscible Gas Drive • Water Flooding
• Gravity Drainage • Horizontal
• Miscible Gas Injection • Vertical
Tertiary Recovery Processes - EOR

• Chemical Flooding
• Polymer
• Surfactant
• Alkali
• Bacterial
• Combinations
• Solvent Flooding
• Liquid Solvents
• Gas Solvents (miscible)
• VAPEX, NSOLV, etc.
Tertiary Recovery Processes - EOR
• Thermal Processes
• Steam Flooding
• Cyclic Steam Stimulation
• Steam Assisted Gravity Drainage
• In-Situ Combustion (Fire Flooding)
• Steam Solvent Combinations
L5A Fractional
Flow Principles
ENPE 525/624
The Buckley-Leverett Theory
• The Buckley-Leverett theory [1942] 𝑞𝑡 = 𝑞𝑜 + 𝑞𝑤
estimates the rate at which an
injected water bank moves
through a porous medium: 𝑞𝑤 = 𝑞𝑡 × 𝑓𝑤
• Flow is linear and horizontal
• Water is injected into an oil reservoir 𝑞𝑜 = 𝑞𝑡 × 𝑓𝑜 = 𝑞𝑡 × 1 − 𝑓𝑤
• Oil and water are both
incompressible
• Oil and water are immiscible
• Gravity and capillary pressure effects
are negligible
The Buckley-Leverett Theory
𝑘𝑟𝑤 𝑘𝐴 𝑑𝑝 𝑘𝑟𝑤 𝐴 𝑑𝑝
𝑞𝑤 = 𝑞𝑤
𝜇𝑤 𝑑𝑥 𝜇𝑤 𝑑𝑥
𝑓𝑤 = =
𝑞𝑜 + 𝑞𝑤 𝑘𝑟𝑜 𝐴 𝑑𝑝 𝑘𝑟𝑤 𝐴 𝑑𝑝
𝑘𝑟𝑜 𝑘𝐴 𝑑𝑝 +
𝜇𝑜 𝑑𝑥 𝜇𝑤 𝑑𝑥
𝑞𝑜 =
𝜇𝑜 𝑑𝑥
𝑘𝑟𝑤
𝑘𝑜 𝜇𝑤 1
• is a function of saturation. So 𝑓𝑤 = =
𝑘𝑤 𝑘𝑟𝑜 𝑘𝑟𝑤 𝑘𝑟𝑜 𝜇𝑤
for constant viscosity fw is just a + 1+
𝜇𝑜 𝜇𝑤 𝑘𝑟𝑤 𝜇𝑜
function of saturation
Fractional
Flow Curve

𝑘𝑟𝑜
= 𝑎𝑒 −𝑏𝑆𝑤
𝑘𝑟𝑤

1
𝑓𝑤 = 𝜇𝑤
1+ ∗ 𝑎𝑒 −𝑏𝑆𝑤
𝜇𝑜
The rate the water enters to the medium element from left hand side (LHS) is:

𝑞𝑡 × 𝑓𝑤 = 𝑤𝑎𝑡𝑒𝑟 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒 𝑒𝑛𝑡𝑒𝑟 𝑡𝑜 𝑡ℎ𝑒 𝑒𝑙𝑒𝑚𝑒𝑛𝑡

The rate of water leaving element from the right hand side (RHS) is:

𝑞𝑡 × 𝑓𝑤 + ∆𝑓𝑤 = 𝑤𝑎𝑡𝑒𝑟 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒 𝑙𝑒𝑎𝑣𝑖𝑛𝑔 𝑡ℎ𝑒 𝑒𝑙𝑒𝑚𝑒𝑛𝑡

The change in water flow rate across the element is found by


Change in Water Flowrate =
performing a mass balance. The movement of mass for an
water entering – water leaving
immiscible, incompressible system gives:
= 𝑞𝑡 × 𝑓𝑤 − 𝑞𝑡 × 𝑓𝑤 + ∆𝑓𝑤
This is equal to the change in element water content per unit
time.
= −𝑞𝑡 × ∆𝑓𝑤
Let Sw is the water saturation of the element at time t. Then if
oil is being displaced from the element, at time (t + Dt) the
water saturation will be 𝑆𝑤 + ∆𝑆𝑤 . So water accumulation in
the element per unit time is:

∆𝑆𝑤 × 𝐴 × 𝜑 × ∆𝑥 ∆𝑆𝑤 × 𝐴 × 𝜑 × ∆𝑥
𝑤𝑎𝑡𝑒𝑟 𝑎𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑡𝑖𝑚𝑒 = = −𝑞𝑡 × ∆𝑓𝑤
∆𝑡 ∆𝑡

∆𝑆𝑤 𝑞𝑡 × ∆𝑓𝑤
In the limit as ∆t → 0 and ∆x → 0 (for the water phase):
=−
∆𝑡 𝐴 × 𝜑 × ∆𝑥

The subscript x on the derivative indicates that this derivative is


different for each element.
It is not possible to solve for the general distribution of water 𝑑𝑆𝑤 𝑞𝑡 𝑑𝑓𝑤
saturation Sw(x,t) in most realistic cases because of the nonlinearity of =−
the problem. For example, water fractional flow is usually a nonlinear 𝑑𝑡 𝑥
𝐴 × 𝜑 𝑑𝑥 𝑡
function of water saturation. It is therefore necessary to consider a
simplified approach
For a given rock, the fraction of flow for water fw is a function only of the water saturation Sw, assuming constant oil and
water viscosities.
The water saturation however is a function of both time and position, which may be express as fw = F(Sw) and Sw = G(t,x).

𝜕𝑆𝑤 𝜕𝑆𝑤 𝑑𝑆𝑤 𝜕𝑆𝑤 𝜕𝑆𝑤 𝑑𝑥


𝑑𝑆𝑤 = 𝑑𝑡 + 𝑑𝑥 = +
𝜕𝑡 𝑥
𝜕𝑥 𝑡
𝑑𝑡 𝜕𝑡 𝑥
𝜕𝑥 𝑡
𝑑𝑡

𝜕𝑥
Now, there is interest in determining the rate of advance of a constant saturation plane, or front , where Sw is
𝜕𝑡 𝑆𝑤
constant and dSw = 0.

𝜕𝑆𝑤
𝑑𝑥 𝜕𝑡 𝑑𝑥 𝑞𝑡 𝑑𝑓𝑤
= 𝑥 =− Buckley-Leverett frontal
𝑑𝑡 𝜕𝑆𝑤
𝜕𝑥
𝑑𝑡 𝑆𝑤
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤
advance equation
𝑡
𝑑𝑓𝑤 𝑑𝑥
The derivative is the slope of the fractional flow curve and derivative is the velocity of the moving plane with
𝑑𝑆𝑤 𝑆 𝑑𝑡 𝑆𝑤
𝑤
water saturation Sw. Because the porosity, area, and flowrate are constant and because for any value of Sw, the derivative
𝑑𝑓𝑤
is a constant, then the rate dx/dt is constant.
𝑑𝑆𝑤 𝑆
𝑤
This means that the distance a plane of constant saturation, Sw, advances is proportional to time and to the value of the
𝑑𝑓𝑤
derivative ( ) at that saturation
𝑑𝑆𝑤
𝑞𝑡 × 𝑡 𝑑𝑓𝑤 𝑄𝑖𝑛𝑗 𝑑𝑓𝑤
𝑥𝑆𝑤 =− =−
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤

Where,

𝑥𝑆𝑤 is the distance traveled by a particular Sw contour

𝑄𝑖𝑛𝑗 is the cumulative water injection at reservoir conditions.


In field units:

5.615 ∗ 𝑞𝑡 × 𝑡 𝑑𝑓𝑤 5.615 ∗ 𝑄𝑖𝑛𝑗 𝑑𝑓𝑤


𝑥𝑆𝑤 =− =−
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤
PROBLEM 1 BUCKLEY LEVERETT ANALYSIS
Assume a cubical reservoir under active water drive with oil production of 900bbl/day. The flow
could be approximated as a linear flow. The cross sectional area is the product of the width, 1320 ft,
and the true formation thickness, 20 ft, so that for a porosity of 0.25, the previous equation becomes

5.615 ∗ 900 × 𝑡 𝑑𝑓𝑤


𝑥𝑆𝑤 =−
0.25 ∗ 1320 ∗ 20 𝑑𝑆𝑤 𝑆𝑤

Consider that because we assume the fluids are completely incompressible, so the oil production rate is equal to
the total flowrate in the different cross sections of the reservoir.
SOLUTION OF PROBLEM 1
If we let x=0 at the first point of the transition zone, then the 𝑑𝑓𝑤 𝑑𝑓𝑤
distances the various constant water saturation planes will 𝑥60 = 46 𝑥120 = 92
𝑑𝑆𝑤 𝑆𝑤
𝑑𝑆𝑤 𝑆𝑤
travel in, say, 60, 120, and 240 days are given by:

The value of the derivative dfwdSw may be obtained for any value of water 𝑑𝑓𝑤
saturation, Sw, by plotting fw from eq. fw versus Sw and graphically taking the slopes at
𝑥240 = 184
𝑑𝑆𝑤 𝑆𝑤
various values of Sw. Assume you find a=1222 and b=12 from Figure slide 9 (intercept
= 1222 = ‘a’ and slope of the straight line = 13 = ‘b’) for eq. fw. For example at Sw = 0.4,
fw = 0.129. The slope taken graphically at Sw = 0.4 and fw = 0.267 is 1.66.
fw does not hold
The derivative dfwdSw may also be obtained mathematically using eq.(4-7):
for the very high
1 and for the quite
𝑓𝑤 = 𝜇𝑤 low water
1+ ∗ 𝑎𝑒 −𝑏𝑆𝑤 saturation ranges
𝜇𝑜

𝜇𝑤 −𝑏𝑆𝑤
𝜕𝑓𝑤 × 𝑏 × 𝑎 × 𝑒 Some small error is
𝜇𝑜
= 2 introduced below
𝜕𝑆𝑤 𝜇𝑤 30% and above 80%
1+ × 𝑎 × 𝑒 −𝑏𝑆𝑤
𝜇𝑜 water saturation
A plot of Sw versus distance the above and typical
fractional flow curves leads to the physically impossible
situation of multiple values of Sw at a given location. For
example Figure shows water saturation distribution
according to eqs. Slide 15

Because of this discontinuity the mathematical approach of


Buckley-Leverett, which assumes that Sw is continuous and
differentiable, will be inappropriate to describe the situation at
the front itself.
The difficulty is resolved by dropping perpendiculars at point Xf
(as flood front position) so that the areas to the right (A) equal
the areas to the left (B), as shown in the figure. In other words
a discontinuity in Sw at a flood front location Xf is needed to
make the water saturation distribution single valued and to
provide a material balance for displacing fluid.
A more elegant method of achieving the same result was
presented by Welge in 1952. This consists of integrating the
saturation distribution over the distance from the injection
point to the front, thus obtaining the average water
saturation behind the front Sw

The situation depicted is at a fixed time, before water


breakthrough, corresponding to an amount of water
injection. At this time the maximum water saturation, Sw = 1
- Sor, has moved a distance X1, its velocity being 𝑄𝑖𝑛𝑗 = 𝑥2 × 𝐴 × 𝜑 × 𝑆𝑤 − 𝑆𝑤𝑐
proportional to the slope of the fractional flow curve
evaluated for the maximum saturation which, as shown in 𝑄𝑖𝑛𝑗 Where, 𝑄𝑖𝑛𝑗 is cumulative
Slide 15, is small but finite. The flood front saturation Swf is 𝑆𝑤 − 𝑆𝑤𝑐 =
𝑥2 𝐴𝜑 water injection
located at position x2 measured from the injection point.
Applying the simple material balance:
𝑄𝑖𝑛𝑗 𝑑𝑓𝑤
Where, 𝑥2 = 𝑥𝑓 = −
𝐴𝜑 𝑑𝑆𝑤 𝑆𝑤
𝑡𝑏𝑡 = Breakthrough time,

𝑞𝑡 = Total injection rate, 𝑄𝑖𝑛𝑗 𝑑𝑓𝑤


𝑥𝑓 = 𝐿 = − , 𝑄𝑖𝑛𝑗 = 𝑞𝑡 × 𝑡𝑏𝑡
𝐴𝜑 𝑑𝑆𝑤
𝐿 = Medium length 𝑆𝑤𝑓
𝐿𝐴𝜑 𝑑𝑓𝑤 𝑞𝑡 × 𝑡𝑏𝑡 The average water saturation in the reservoir at
= , = 𝑃𝑉𝐼 𝑏𝑡
𝑞𝑡 𝑡𝑏𝑡 𝑑𝑆𝑤 𝑆𝑤𝑓
𝐴𝜑𝐿 the time of breakthrough is given by material
balance as:
Where PVI is the pore volume injected.
𝑆𝑤𝑏𝑡 = 𝑆𝑤𝑖 + 𝑃𝑉𝐼 𝑏𝑡
1 𝑑𝑓𝑤
=
𝑃𝑉𝐼 𝑏𝑡 𝑑𝑆𝑤 𝑆𝑤𝑓 1
𝑆𝑤𝑏𝑡 − 𝑆𝑤𝑖 = 𝑃𝑉𝐼 𝑏𝑡 =
𝑑𝑓𝑤
𝑆𝑤𝑏𝑡
𝑑𝑆𝑤 𝑆𝑤𝑓

Swf, fwf

𝑑𝑓𝑤 1
=
𝑑𝑆𝑤 𝑆𝑤𝑓 𝑆𝑤𝑏𝑡 − 𝑆𝑤𝑖
𝑥
𝑥𝐷 =
𝐿

𝑞𝑡 𝑡
𝑡𝐷 =
𝐴𝜑𝐿

𝑥𝐷 = Normalized distance
𝑥𝐷𝑆𝑤 = 𝑡𝐷 𝑓𝑆′𝑤
𝑡𝐷 = Pore volumes injected
JBN Analysis

• Unsteady state displacements analysis is described by Johnson, Bossler and Neumann, which is summarized below.
Three calculation stages are involved:

o The ratio kro/krw.


o The values of kro and hence krw.
o The value of Sw.
o The method is aimed at giving the required values at the outlet face of the core which is essentially where
volumetric flow observations are made.
(a). kro/krw:
• The average water saturation (Swav) is plotted against Qi (Figure 2-91-a). Fractional flow of oil, at the outlet face of
the core sample is:

𝑑𝑆𝑤𝑎𝑣
𝑓𝑜 =
𝑑𝑄𝑖
𝑘𝑟𝑜 𝑘𝐴 ∆𝑃
𝑄𝑜 𝑄𝑜 𝜇𝑜 𝐿 1
𝑓𝑜 = = = =
𝑄𝑡 𝑄𝑤,𝑖𝑛𝑗 𝑘𝑟𝑤 𝑘𝐴 ∆𝑃 𝑘𝑟𝑤 𝜇𝑜
1 +
𝜇𝑤 𝐿 𝑘𝑟𝑜 𝜇𝑤

(a) (b)
(a) Average Water saturation vs. Water Injection, (b) Injectivity Ratio
(b). kro:
• A plot of ∆P/∆Pi against Qi is used to obtain injectivity ratio (Figure 2-91-b).
∆𝑃𝑖 1
𝐼𝑅 =
∆𝑃 𝑄𝑖
1 1
kro is obtained by plotting versus :
𝑄𝑖𝐼𝑅 𝑄𝑖

and using the relationship:

1
𝑘𝑟𝑜 = 𝑓𝑜
𝑑 1Τ𝑄𝑖 𝐼𝑅
𝑑 1Τ𝑄𝑖
krw can be calculated from eq. (2-172), or:
1 − 𝑓𝑜 𝜇𝑤
𝑘𝑟𝑤 = 𝑘
𝑓𝑜 𝜇𝑜 𝑟𝑜
L5B Mobilizing
Trapped Oil
ENPE 525/624
Trapping of Oil in Porous Media

 1 1 
DP = PB − PN = 2 ow  − 
 R1 R 2 

(Source: McAuliffe, 1973)

• DP is called the “Jamin Effect” – this is the Pc differential pressure


that must be overcome to mobilize oil droplets surrounded by water
Trapping of Oil, cont.

 1 1 
DP = PB − PN = 2 ow  − 
 R1 R 2 

Interfacial tension (mN/m) Pressure gradient (kPa/m)


14 11,670 (516 psi/ft) Assume R for 12 D rock, with
1.4 1167 (52 psi/ft) R1 = 0.5R and R2 = 1.5R
0.14 117 (5 psi/ft)
0.014 11.7 (0.5 psi/ft)

• Actual pressure gradients within reservoir can vary from 5 – 25 kPa/m


• The only way that trapped oil can be mobilized is if IFT is reduced by
several orders of magnitude
Concept of Pore Doublet Model

• Under dominance of capillary forces, water moves


through smaller tube, traps oil in larger tube
• Once water moves through doublet, oil is
discontinuous (trapped)
Capillary Number

• Dimensionless term
• Ratio of viscous forces to capillary forces
• During the solution of flow through each line of the pore doublet, this
term was observed to be present
• Trapping of discontinuous oil is related to this ratio of forces
No. Author(s) Year Porous Media Correlating Group

1 Fairbrother and Stubbs 1935 Capillary 𝑣𝜇Τ𝜎

2 Leverett 1939 Sandstone 𝐿𝑃𝑐 Τ 𝐷∆𝑃

3 Brownell and Katz 1947 Sandstone 𝐾∆𝑃 Τ 𝑔 𝑐𝑜𝑠 𝜃 𝐿𝜎 According to the work of
4 Ojeda, Preston and Calhoun 1953 Sandstone 𝜎Τ∆𝑃 Chatzis and Morrow (1984) the
5 Moore and Slobod 1956 Sandstone 𝑣𝜇 Τ 𝜎 𝑐𝑜𝑠 𝜃 range of capillary numbers over
6 Saffman and Taylor 1958 Hell-Shaw Cell 𝑣𝜇 Τ𝜎 which capillary displacement is
7 Taber 1969 (Berea) Sandstone ∆𝑃 Τ 𝐿𝜎 dominant is Nca < 10-5 - 10-4.
8 Foster 1973 (Berea) Sandstone 𝑣𝜇 Τ 𝜎𝜑 When the capillary number
9 Lefebvre duPrey 1973 Teflon, Steel, and Aluminum 𝜎Τ 𝑣𝜇
exceeds a value of 10-4 then the
10 Melrose and Brandner 1974 Unconsolidated Glass Beads 𝑣𝜇 Τ 𝜎𝜑
residual oil is mobilized through
11 Ehrlich, Hasiba and Raimondi 1974 Sandstone 𝐾∆𝑃 Τ 𝜎𝜑𝐿
a stripping process.
12 Abrams 1975 Sandstone, Limestone 𝑣𝜇1.4 0.4
𝑤 Τ 𝜎𝜇𝑜

13 Reed and Healy 1977 Various 𝐾∆𝑃 Τ 𝜎𝐿𝑐𝑜𝑠 𝜃


Criteria for Enhanced Oil Recovery

(Source: Moore and Slobod, 1956)

(Source: Moore and Slobod, 1956)

• As capillary forces (IFT) are reduced, more oil can be recovered


Understanding Trapping of Oil

• Lower residual oil saturation at higher Nca values


• Higher viscous forces/lower capillary forces lead to mobilization of trapped oil
droplets
• Assumptions that we are making:
• Water wet media: we are mobilizing discontinuous oil!
• Viscosity = displacing phase (water) value; this equation assumes similar oil and
water viscosity
• NOTE: there is no impact of oil viscosity in Nca!
Impact of Oil Viscosity on Trapping

(Source: Abrams, 1974)


Impact of Oil Viscosity on Trapping, cont.

(Source: Abrams, 1974)


Understanding Capillary Number and Oil Trapping

• Capillary number does not consider properties of oil; only considers


mobilization of trapped oil
• When oil is more viscous than water, we also need to understand the
mobility ratio of the system
• For M > 1, water will finger through the oil
• Oil is CONTINUOUS, not trapped like for Nca analysis
Effect of Oil Viscosity on Displacement
NOTE:
- Low oil recovery is not due to
trapping of discontinuous oil.
- Instead, oil is continuous but
bypassed by water.
- To recover this oil, the goal is
improved sweep, not
reduction in capillary forces!

(Source: Craig, 1971)


Reservoir Heterogeneity and Oil Recovery

𝑄𝑓𝑟𝑎𝑐 𝑘𝑓𝑟𝑎𝑐 Δ𝑃
= 𝑣𝑒𝑙𝑓𝑟𝑎𝑐 =
𝐴𝑓𝑟𝑎𝑐 𝜇 𝐿

𝑄𝑚𝑎𝑡 𝑘𝑚𝑎𝑡 Δ𝑃
= 𝑣𝑒𝑙𝑚𝑎𝑡 =
𝐴𝑚𝑎𝑡 𝜇 𝐿

(Source: Warren and Root, 1971)

• Permeability of fractures >> matrix permeability


• Injection water goes through fractures, residual oil continuous in matrix but
surrounded by fractures filled with water
Oil Production in Heavy Oil Primary Production

(Source: www.onepetro.org)

• Production of oil with sand with oil leads to the formation of very high permeability
channels (wormholes) in sand
• Similar to fractured reservoir case, but wormholes cover an unknown extent
Injection in Wormhole Heavy Oil Systems

• Wormhole permeability >> sand permeability


• Any injection fluid will travel through wormholes
• Oil left in remaining sand is continuous but bypassed due to permeability (plus high oil viscosity)
• Recovery of oil requires finding a way to access the bypassed oil, not mobilization of trapped oil
(Nca analysis does not make sense here!)
Understanding EOR Potential

• Nca deals with mobilization of discontinuous trapped oil in water wet


porous media
• With higher Nca, residual oil saturation is reduced
• Concept: reduce capillary trapping of oil (surfactants and polymers)
• Sometimes oil is not trapped by capillary forces, but instead is
bypassed due to high oil viscosity or variable sand permeability
• In this case, EOR is not related to Nca, but instead focuses on improving the
mobility ratio (sweep efficiency) in the system

You might also like