Department of Petroleum Engineering and Applied Geophysics

Examination paper for TPG4160 Reservoir

Simulation

Academic contact during examination: Jon Kleppe

Phone: 91897300/73594925

Examination date: June 2, 2016

Examination time (from-to): 9:00-13:00
Permitted examination support material: D/No printed or hand-written support
material is allowed. A specific basic calculator is allowed.

Other information:

Language: English
Number of pages: 11
Number of pages enclosed: 0

Checked by:

____________________________
Date Signature
Final Exam page 2 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 1 (5 points)

This question relates to the Gulltopp Field project work.

a) The reservoir pressure of Gulltopp was slowly declining before production start in 2008.
Why?
b) Was the amount of outflow from Gulltopp before production start in 2008 significant
compared to the total produced volumes until 2014? Why or why not?
c) What was the main driving mechanism during production of Gulltopp?
d) What were the main uncertainties in the simulation results?
e) Discuss briefly how oil recovery from Gulltopp could have been improved using
alternative development strategies.

Solution

Any answer that shows that they have been working actively with the group project

Question 2 (10 points)

List all steps and standard relationships/formulas involved in deriving partial differential flow
equations for flow in porous media. Black-Oil, one-phase, one-dimensional, horizontal flow is
sufficient.

Solution

a) Starting with a mass balance for the control element:

⎧Mass into the ⎫ ⎧Mass out of the ⎫ ⎧Rate of change of mass ⎫

⎨ ⎬− ⎨ ⎬= ⎨ ⎬
⎩element at x ⎭ ⎩element at x+Dx ⎭ ⎩inside the element ⎭
b) then substitute for the relationship between velocity and pressure
k ∂P
• Darcy’s: u = − ,
µ ∂x
∂P µ
• or Forchheimer’s (for high velocity flow): − = u + βun ,
∂x k
• or Brinkman’s (applies to both porous and non-porous flow):
∂P µ ∂ 2u
− =u −µ 2 .
∂x k ∂x
c) Thirdly introducing the relationship between porosity and pressure:
1 ∂φ
c r = ( )( ) T
φ ∂P
d) and finally the relationship between fluid density and pressure:
ρ o S + ρ g sRs o constant
• Oil density: ρo = , which for one-phase flow becomes: ρ o =
Bo Bo
P ZS
• Real gas equation: PV = nZRT , which leads to: ρg = ρ g S
Z PS
Final Exam page 3 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 3 (10 points)

List all steps and standard approximations for converting a continuous partial differential
equation to discrete form. Include a sketch of the continuous and the discrete (gridded) flow
system. Include standard approximations needed for the simple diffusivity equation:
∂2 P φµc ∂P
2 =( )
∂x k ∂t

Solution

a) Continuous system:

∂2 P φµc ∂P
= ( )
∂x2 k ∂t
b) Discrete system:
1 i-1 i i+1 N

Δx

t+Δt
Pi+1 − 2Pi t+Δt + Pi−1
t+Δt
φµc Pi t+Δt − Pi t
≈( ) , i = 1,..., N
Δx 2 k Δt

c) Taylor approximations:
−Δt (−Δt)2
P(x,t) = P(x,t + Δt) + P ′(x,t + Δt) + P ′′(x,t + Δt) + .....
1! 2!
leading to:

∂ P t+Δt Pi t+Δt − Pi t
( ) = + O(Δt) .
∂t i Δt
Δx (Δx)2
P(x + Δx,t + Δt) = P(x,t + Δt) + P ′(x,t + Δt) + P ′′(x,t + Δt) + .....
1! 2!

(−Δx) (−Δx)2
P(x − Δx,t + Δt) = P(x,t + Δt) + P ′(x,t + Δt) + P ′′(x,t) + .....
1! 2!

leading to:

∂ 2 P t+Δt Pi+1
t+Δt
− 2Pi t+Δt + Pi−1
t+Δt
( 2 )i = + O(Δx 2 ) .
∂x (Δx) 2
Final Exam page 4 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 4 (9 points)
The discretized form of the left hand side of the oil equation may be written in terms of
transmissibility and pressure differences, as:

∂ ⎛ k kr o ∂Po ⎞
≈ Txoi +1/ 2 (Po i+1 − Po i ) + Txoi −1/ 2 (Po i−1 − Po i )
∂ x ⎜⎝ µo Bo ∂ x ⎟⎠ i
Using the following transmissibility as example,
2ki+1/2 λoi+1/2
Txoi+1/2 =
Δxi ( Δxi+1 + Δxi )

a) What is the averaging method normally applied to absolute permeability between grid
blocks ( ki+1/2 )? Why? Write the expression for average permeability between grid blocks
(i+1) and (i).
b) Write an expression for the selection of the conventional upstream mobility term ( λoi+1/2 )
for use in the transmissibility term of the oil equation above for flow between the grid
blocks (i+1) and (i).
c) Make a sketch of a typical Buckley-Leverett saturation profile resulting from the
displacement of oil by water (i.e., analytical solution). Then, show how the corresponding
profile, if calculated in a numerical simulation model, typically is influenced by the
choice of mobilities between the grid blocks (Sketch curves for saturations computed
with upstream or average mobility terms, respectively).

Solution

a) Harmonic average is used because it properly represents flow in series across blocks of
different permeabilities. It may be derived from Darcy's law (steady flow).
Δxi + Δxi+1
ki+1/2 =
Δxi Δxi+1
+
ki ki+1
⎧⎪ λ oi+1 if Poi+1 ≥ Poi
λ =
b) i+1/2 ⎨
o
⎪⎩ λ oi if Poi+1 < Poi
c)

Qw
Sw
1-Swir
exact
average
upstream
S wir

x
Final Exam page 5 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 5 (12 points)

For two-phase flow of oil and water in a horizontal, one-dimensional porous medium, the flow
equations can be written (including well terms):
∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φSo ⎞
⎜ ⎟ − q′o = ⎜ ⎟
∂x ⎝ µo Bo ∂x ⎠ ∂t ⎝ Bo ⎠
∂ ⎛ kkrw ∂Pw ⎞ ∂ ⎛ φS w ⎞
⎜ ⎟ − q′w = ⎜ ⎟,
∂ x ⎝ µw Bw ∂x ⎠ ∂t ⎝ Bw ⎠
where
Pw = Po − Pcow
So + Sw = 1
a) Write the two flow equations on discretized forms in terms of transmissibilities, storage
coefficients and pressure and saturation differences (Do not derive).
b) List the assumptions for IMPES solution, and outline briefly how we solve for pressures
and saturations
c) What are the limitations of the IMPES solution?

Solution

Txoi+1 2 ( Poi+1 − Poi ) + Txoi−1 2 ( Poi−1 − Poi ) − q′oi

a)
= Cpooi ( Poi − Poit ) + Cswoi ( Swi − Swit ), i = 1,N

Txwi+1 2 [( Poi+1 − Poi ) − ( Pcowi+1 − Pcow i )] + Txwi−1 2 [( Poi−1 − Poi ) − ( Pcowi−1 − Pcow i )] − q′wi
= Cpowi ( Poi − Poit ) + Cswwi ( Swi − Swit ), i = 1,N
b) ALL DETAILS BELOW ARE NOT NEEDED, BUT A PROPER OUTLINE IS
REQUIRED. In the IMPES solution, all coefficients and capillary pressures are evaluated at
time=t. The two equations are combined so that the saturation terms are eliminated. The
resulting equation is the pressure equation:
ai Poi−1 + bi Poi + c i Poi+1 = di , i = 1,N
which may be solved for pressures in all grid blocks by Gaussian Elimination Method, or some
other method. Then, the saturations may be solved for explicitly by using one of the equations.
Using the oil equation, yields:
Swi = Swit +
1
t
Cswoi
[
t
Txoi+1/ 2 ( Poi+1 − Poi ) + Txoi−1/ 2 ( Poi−1 − Poi ) − q′oi − Cpooi ( Poi − Poi )
t t t
]
, i = 1,N

c) The approximations made in the IMPES method, namely the evaluation of coefficients at
old time level when solving for pressures and saturations at a new time level, puts restrictions
on the solution which sometimes may be severe. Obviously, the greatest implications are on
the saturation dependent parameters, relative permeability and capillary pressure. These
change rapidly with changing saturation, and therefore IMPES may not be well suited for
problems where rapid variations take place.

IMPES is mainly used for simulation of field scale systems, with relatively large grid blocks
and slow rates of change. It is normally not suited for simulation of rapid changes close to
wells, such as coning studies, or other systems of rapid changes.

However, provided that time steps are kept small, IMPES provides accurate and stable
solutions to a long range of reservoir problems.
Final Exam page 6 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 6 (12 points)

For a one-dimensional, horizontal, 3-phase oil, water, gas system, the general flow equations
are (including well terms):
∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φSo ⎞
⎜ ⎟ − q′o = ⎜ ⎟,
∂x ⎝ µo Bo ∂x ⎠ ∂t ⎝ Bo ⎠
∂ ⎛ kk rg ∂Pg kk ∂P ⎞ ∂ ⎛ φS φS ⎞
⎜⎜ + Rso ro o ⎟⎟ − q′g − Rsoq′o = ⎜⎜ g + Rso o ⎟⎟
∂x ⎝ µg Bg ∂x µo Bo ∂x ⎠ ∂t ⎝ Bg Bo ⎠
∂ ⎛ kk rw ∂Pw ⎞ ∂ ⎛ φSw ⎞
⎜ ⎟ − qw′ = ⎜ ⎟
∂x ⎝ µw Bw ∂x ⎠ ∂t ⎝ Bw ⎠
a) Explain briefly the physical meaning of each term in all three equations.
b) What are the criteria for saturated flow? What are the functional dependencies of
Rso and Bo ?
c) What are the primary unknowns when solving the saturated equations?
d) What are the criteria for undersaturated flow? What are the functional dependencies of
Rso and Bo ?
e) What are the primary unknowns when solving the undersaturated equations?
f) Rewrite the equations above for undersaturated flow conditions.

Solution
∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φS o ⎞
a) ⎜ ⎟− q′o = ⎜ ⎟
∂x ⎝ µo Bo ∂x ⎠ ∂t ⎝ Bo ⎠
transport of oil oil prod. accumulation of oil
∂ ⎛ kk rg ∂Pg kk ∂P ⎞ ∂ ⎛ φSg φSo ⎞
⎜⎜ + Rso ro o ⎟⎟ − q′g − Rsoq′o = ⎜ + Rso ⎟
∂x ⎝ µg Bg ∂x µo Bo ∂x ⎠ ∂t ⎜⎝ Bg Bo ⎟⎠
transport of transport of free gas prod. sol. gas prod accumulation. accumulation
free gas sol. gas of free gas of solution gas
∂ ⎛ kk rw ∂Pw ⎞ ∂ ⎛ φSw ⎞
⎜ ⎟− q′w = ⎜ ⎟
∂x ⎝ µw Bw ∂x ⎠ ∂t ⎝ Bw ⎠
transport of water water prod. accumulation of water

b) Saturated flow: criteria Po = Pbp and Sg > 0.

dependencies Bo = f ( Po ) and Rso = f ( Po ) .
c) Po and Sg
d) Undersaturated flow: criteria Po > Pbp and Sg = 0.
dependencies Bo = f ( Po ,Pbp ) and Rso = f ( Pbp ) .
e) Po and Pbp

∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φSo ⎞
f) ⎜ ⎟ − q′o = ⎜ ⎟
∂x ⎝ µo Bo ∂x ⎠ ∂t ⎝ Bo ⎠
∂⎛ kkro ∂Po ⎞ ∂ ⎛ φSo ⎞
⎜ Rso ⎟ − q′g − Rsoq′o = ⎜ Rso ⎟,
∂x ⎝ µo Bo ∂x ⎠ ∂t ⎝ Bo ⎠
Final Exam page 7 of 11
TPG4160 Reservoir Simulation, June 2, 2016

∂ ⎛ kk rw ∂Pw ⎞ ∂ ⎛ φSw ⎞
⎜ ⎟ − q′w = ⎜ ⎟
∂x ⎝ µw Bw ∂x ⎠ ∂t ⎝ Bw ⎠

Question 7 (12 points)

For two-phase flow (constant flow area) the right hand side of the oil and gas equations may
be written (saturated oil case):
∂ ⎛ φ So ⎞
∂ t ⎜⎝ Bo ⎟⎠
∂ ⎛ φ Sg φS ⎞
⎜ + Rso o ⎟
∂ t ⎝ Bg Bo ⎠
The corresponding discretized forms are:
( ) (
Cpooi Poi − Poit + Csgoi Sgi − Sgit )
Cpogi (P oi − Poit ) + C (S
sggi gi − Sgit )
a) Sketch typical curves that show the pressure dependencies of Bo , Bg, Rso .
b) Show the complete derivations of the four coefficients Cpooi , Csgoi , Cpogi , Csggi .

Solution
a)

b) Students should show complete derivations

φi (1 − Sg i ) ⎡ cr d (1 / Bo ) ⎤
C pooi = ⎢ B + dP ⎥
Δt ⎣ o o ⎦i

φi
C sgoi = −
BoiΔti

φ i ⎡ ⎛ c r d(1 / Bg ) ⎞ ⎛c d(1 / Bo ) ⎞ (1 − Sg ) dRs o ⎤

C pogi = ⎢ Sg ⎜ + ⎟ + Rs o (1 − Sg )⎜ r + + ⎥
Δt ⎢⎣ ⎝ Bg dPg ⎠ ⎝ Bo dPo ⎟⎠ Bo dPo ⎥⎦
i

φi ⎡ ⎛ c r d(1/ Bg ) ⎞ dPcog Rs o 1 ⎤
C sggi = ⎢S ⎜ + ⎟ − + ⎥
Δt ⎢⎣ g ⎝ Bg dPg ⎠ dSg Bo Bg ⎥⎦
i
Final Exam page 8 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Question 8 (10 points)

Normally, we use either a Black Oil fluid description or a compositional fluid description in
reservoir simulation.

a) What are the components and the phases used in Black Oil modeling?
b) What are the components and the phases used in compositional modeling?
c) Write the standard flow equations for the components required for Black Oil modeling
(one dimensional, horizontal, constant flow area).
d) Write the standard flow equations the components required for compositional modeling
(one dimensional, horizontal, constant flow area). Let
C kg = mass fraction of component k present in the gas phase
C ko = mass fraction of component k present in the oil phase.
e) A Black Oil fluid description may be regarded as a subset of a compositional fluid
description. Define the pseudo-components required in order to reduce the compositional
equations to Black Oil equations (one-dimensional, horizontal, constant flow area)

Solution

a) Components: oil and gas, phases: oil and gas

b) Components: hydrocarbons ( C1 H 4 , C2 H 6 , C3 H 8 ,.... ) and non-
hydrocarbons ( CO2 , H 2 S, C2 ,.... ) , phases: oil and gas
c)
∂ ⎛ kk rg ∂Pg kkro Rso ∂Po ⎞ ∂ ⎡ ⎛ Sg So Rso ⎞⎤
⎜ + ⎟ = ⎢φ⎜ + ⎟⎥
∂x ⎜⎝ Bg µg ∂x Boµo ∂x ⎟⎠ ∂t ⎢⎣ ⎜⎝ Bg Bg ⎟⎠⎥⎦
∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φSo ⎞
⎜ ⎟= ⎜ ⎟
∂x ⎝ Boµo ∂x ⎠ ∂t ⎝ Bo ⎠

d)
∂⎛ kk ∂P kk ∂P ⎞ ∂
⎜⎜Ckg ρ g rg g + Cko ρ o ro o ⎟⎟ =
∂x ⎝ µ g ∂x µ o ∂x ⎠ ∂t
[ ]
φ (Ckg ρ g Sg + Cko ρ o So ) , k = 1,N c

e)
The Black Oil model may be considered to be a pseudo-compositional model with two
components. Define the components and the fractions needed to convert the
compositional equations to Black-Oil equations.

component 1: oil – k=o

∂⎛ kk ∂P kk ∂P ⎞ ∂
⎜⎜Cog ρ g rg g + Coo ρ o ro o ⎟⎟ =
∂x ⎝ µg ∂x µ o ∂x ⎠ ∂ t
[
φ (Cog ρ g Sg + Coo ρ o So ) ]
component 2: gas – k=g
Final Exam page 9 of 11
TPG4160 Reservoir Simulation, June 2, 2016

∂⎛ kk ∂P kk ∂P ⎞ ∂
⎜⎜Cgg ρ g rg g + Cgo ρ o ro o ⎟⎟ =
∂x ⎝ µg ∂x µ o ∂x ⎠ ∂ t
[ ]
φ (Cgg ρ g Sg + Cgo ρ o So )

Question: what are the fractions needed to get the Black Oil equations:
∂ ⎛ kk rg ∂Pg kkro Rso ∂Po ⎞ ∂ ⎡ ⎛ Sg So Rso ⎞⎤
⎜ + ⎟ = ⎢φ⎜ + ⎟⎥
∂x ⎜⎝ Bg µg ∂x Boµo ∂x ⎟⎠ ∂t ⎢⎣ ⎜⎝ Bg Bg ⎟⎠⎥⎦
∂ ⎛ kk ro ∂Po ⎞ ∂ ⎛ φSo ⎞
⎜ ⎟= ⎜ ⎟
∂x ⎝ Boµo ∂x ⎠ ∂t ⎝ Bo ⎠
Answer:

fraction of "gas in gas": Cgg = 1

fraction of "oil in gas": Cog = 0
ρ gS Rso
fraction of "gas in oil": Cgo =
ρ o Bo
ρ
fraction of "oil in oil": Coo = oS
ρ o Bo

Question 9 (10 points)

Normally, we use either a conventional single porosity model or a fractured dual porosity
model model in simulation of a reservoir.

a) Describe the main differences between a conventional reservoir and a fractured reservoir,
in terms of the physics of the systems.
b) How can we identify a fractured reservoir from standard reservoir data?
c) Explain briefly the primary concept used in deriving the flow equations for a dual-
porosity model.
d) Write the basic equations (one-phase, one-dimension) for
• a two-porosity, two-permeability system
• a two-porosity, one-permeability system
e) In terms of the physics of reservoir flow, what is the key difference between the two
formulations in question d)?
f) How is the fluid exchange term in the flow equations in question d) represented? What are
the shortcomings of this representation?

Solution

a) Conventional: One porosity, one permeability system, with one flow equation for each
component flowing.
Fractured: Two porosities, two permeabilities system, withmost of the fluids in the
matrix system, and most of the transport capacity in the fracture system. Requires two
flow equations for each component flowing.
b) K core  K welltest
c) The matrix system supplies fluids to the fracture system, by whatever mechanisms
present (depletion, gravity drainage, imbibition, diffusion,...), and the fracture system
Final Exam page 10 of 11
TPG4160 Reservoir Simulation, June 2, 2016

transports the fluids to the wells. Som transport may also occure in the matrix system,
from block to block, provided that there is sufficient contact.
d) Dual porosity, dual permeability model:

∂ ⎛ k ∂P ⎞ ∂ ⎛φ⎞
⎜ ⎟ + qmf
′ = ⎜ ⎟
∂ x ⎝ µ B ∂x ⎠ f ∂t ⎝ B ⎠ f

∂ ⎛ k ∂P ⎞ ∂ ⎛φ⎞
⎜ ⎟ − qmf
′ = ⎜ ⎟
∂ x ⎝ µ B ∂x ⎠ m ∂t ⎝ B ⎠ m

One porosity, one permeability model (fracture eqn.):

∂ ⎛ k ∂P ⎞ ∂ ⎛φ⎞
⎜ ⎟ + qmf
′ = ⎜ ⎟
∂ x ⎝ µ B ∂x ⎠ f ∂t ⎝ B ⎠ f

Conventional fracture models represents the exchange term by

(
′ = σλ Pm − Pf
− qmf )
where σ is a geometric factor, λ is the mobility term, and Pm and Pf represent matrix
and fracture pressures, respectively.
e) In the dual porosity, dual permeability model, fluid flow may occur11 from one matrix
block to another. In the one porosity, one permeability model, all flow occurs in the
fracture system
f) The exchange term is conventionally defined as

(
′ = σλ Pm − Pf
− qmf )
where σ is a geometric factor, λ is the mobility term, and Pm and Pf represent matrix
and fracture pressures, respectively.
Obviously, this term cannot adequately represent the flow mechanisms present, such
as depletion, gravity drainage, imbibition, diffusion,... In addition, an average pressure
for the matrix block is used in the exprexxion, so that pressure gradients inside the
block is not accounted for.

Question 10 (10 points)

For a one-dimensional, vertical (z), 3 phase oil, water, gas system, outline how initial pressures
and saturations may be computed in a simulation model, assuming that equilibrium conditions
apply:
a) Sketch the reservoir, with a grid superimposed, including gas-oil-contact (GOC) and water-
oil-contact (WOC).
b) Sketch the oil-gas and oil-water capillary pressure curves, and show the how the initial
equilibrium pressures and saturations are determined in the continuous system.
c) Sketch the initial saturations as they are applied to the grid blocks.
Final Exam page 11 of 11
TPG4160 Reservoir Simulation, June 2, 2016

Solution