ARS-Day1 1of 3

Applied Reservoir Simulation
Schlumberger Public
Course
Reservoir Simulation Application Training

Course
and (Eclipse) Workshop
GeoQuest Training and Development,
Denver and Houston
March 06 Applied Reseervoir Simulation Day 1 1

Course Objectives
– Understand The Basic Principles and Basic
Equations Involved in Reservoir Simulation.
– Understand The Data Requirements For
Conducting A Study.
Schlumberger Public
– Understand Current Practice In Using Coarse
And Finely Gridded Models To Incorporate
Heterogeneity.
– Understand The Simulation Study Approach
Which Leads To A Quality Result.
– Understand The Different Types Of Models
Available.

Course Objectives
– Understand and practice History
Matching
– Understand the errors present in
Schlumberger Public
numerical simulators
– Understand scale-up and
Pseudofunctions
– See an overview of fractured reservoir
simulation or streamline simulation and
compositional simulation

Workshop Objectives
• Demonstrate key factors in reservoir

simulation by having students run
simulation cases and compare
Schlumberger Public
results.
• Have students use reservoir
simulator as a tool to investigate
various reservoir engineering /
reservoir management principles

Outline of Lecture Topics
• Day 1
– Introduction
• Course Objectives
Schlumberger Public
• Simulation Process
– Classical Analysis
– Reservoir Simulation Overview
– Overview of the Workflow
– Characterizing the Reservoir
– Benefits of Reservoir Simulation

Outline of Lecture Topics
• Day 1
– Formulation of Equations
– Linearization and Solution Process
Schlumberger Public
– Minimum Data Requirements
– Well Calculations

Outline of Workshop Problems
• 1. Problem 1:
A. IMPES and Implicit Comparison
B. Time Truncation Tests
Schlumberger Public
C. Numerical Dispersion

• 2. Problem 2:
A. Water Coning – Critical Coning Rate
B. Water Influx – History Matching Kh
Schlumberger Public
C. Water Coning
I. History Matching Kv
II. Creation of Pseudo Krw in Coarse Grid to
Match Coning
D. Vertical Equilibrium Comparison

• 3. Problem 3: History Matching

Primary Production
A. Sensitivity Simulations
Schlumberger Public
B. History Match
C. Predictions

INTRODUCTION
Schlumberger Public
Old and New Names
• New Names for the ECLIPSE

Simulators:
• ECLIPSE 100 = ECLIPSE Black Oil
Schlumberger Public
• ECLIPSE 300 = ECLIPSE
Compositional
• ECLIPSE 500 = ECLIPSE Thermal

The Simulation Process
Mathematical Model
Schlumberger Public
Engineering/Simulation
∂  kkrp ∂Φ  ∂  kkrp ∂Φ p 
+ +
Model
∂x  µ p B p ∂x  ∂y  µ p B p ∂y 
∂  kkrp ∂Φ p  ∂  φSp 
=
∂z  µ p B p ∂z  ∂t  B p 

Shared Earth Model
Schlumberger Public
Reservoir Characterization Efforts
in View of Reservoir Simulation
Seismic Outcrop Core Pore Scale
o o o o
o o o o
o o o
o o o o o o o
o o o o
o o o o o o o
o o o o
Schlumberger Public
Well test Well Logs
Data Upscaling/Downscaling
Integration
Finite Element Mesh ,t)

P(r
e
P(r ,t)
w
i
i+1
Finite Difference Mesh

March 06 Applied Reseervoir Simulation
for Day 1
Macro-Flow 14
Schlumberger Public
Production System
Schlumberger Public
Production System
Facilities
Pipelines Separator,
pump,
• Network interactions
compressor, etc.
• Slugging
• Thermal performance
Schlumberger Public
Riser
Flowline (buried,
Choke insulated etc.)
Wells
• Completion design
Gas lift or ESP Surface
• Artificial lift
• Well performance Pipeline
Tubing or Network
Annular
Deviated well
Flow
Horizontal &
Multilateral
Completions
Total Production Compress
System Separator or
Ris Pum
Choke
er p gas
Schlumberger Public
Flowline
oil
Safety Export
Valve lines
Tubing
Reservoi
Completion r
Integrated Model
Schlumberger Public
Purpose of Reservoir Simulation
and Reservoir Management
with reservoir
simulation and
reservoir
management
Schlumberger Public
Oil Production Rate
without
reservoir
simulation and
reservoir
management time
today

Reservoir Simulation
• What is it ?
– numerical model of reservoir made up
of a large number of cells. Equations
Schlumberger Public
are solved to calculate pressures and
flows
• What can it do ?
– used to predict future performance to
decide on optimum development
strategies

• What data does it need

– structural information, rock properties,
fluid properties, well data, historical
Schlumberger Public
production and operating constraints
• Underlying concepts
– Mass Balance
– Darcy’s Law
– Well inflow model
– Physical PVT model

SIMULATION:
Schlumberger Public
``To Give the Appearance Of...''

Reservoir Models Used: History of
Simulation
• Analogy - Well Productivity
- Recovery Factors
- Reservoir Data
• Decline Analysis - History Match
- No Operational Change
Schlumberger Public
- Indication of Drive Mechanism
• Classical Analysis - Volumetric Analysis
- Material Balance
- Fluid Displacement
• Physical Models - Resistance/Capacitance
- Analogue
- Potentiometric
• Numerical Models - Streamtube
- Finite Element
- Finite Difference

Overview of Classical Analysis
Schlumberger Public
Classical Analysis
• AREAS OF ANALYSIS :
– VOLUMETRIC
Schlumberger Public
– MATERIAL BALANCE
– FLUID DISPLACEMENT

Limitations of Classical Methods
– NO DISTRIBUTION OF PRESSURES
AND SATURATIONS
– NO INDICATION OF UNSWEPT OR
Schlumberger Public
UNCONTACTED AREAS
– DOES NOT USE SPATIAL
INFORMATION

Volumetric Calculations
• VOLUMETRIC CALCULATIONS :
Schlumberger Public
OIP= AREA * hnet * Φ* (1 - Sw ) / Bo

Material Balance
• 2. MATERIAL BALANCE :
RELATES PRESSURE/PRODUCTION with the

energy in the system
Schlumberger Public
Gas Cap Expansion
Solution Gas Expansion
Remaining Oil
Rock & Connate Water Expansion
Net Water Influx Production Injection

Drive Types
• Fluid Expansion: Press > Pb
Solution Gas/Secondary Gas Cap:

Press < Pb. Gravity Drainage
Schlumberger Public
Primary Gas Cap: Pres < Pb
Compaction Drive: Unconsolidated sands,

over pressured conditions
Water Drive: Aquifer influx due to

expansion or high pressure, or water
injection
Process Included in GMBE
G ENE RAL M AT ERIAL BA LANCE E Q UATIO N

(G M BE )
Gas
Schlumberger Public
Injec tion O IL +GA S +WA T ER
PR O D U CT IO N
Ga s
Oil
Wate r
Gas
Oil Ap pro xim at es res ervo ir

a s a t an k w it h g ive n
W ater b ou n da ries an d
u nifo rm T a n d P.
D rive
W ater M B C a lcu lat io n is a
ve ry g ross
W ater a pp ro xim a tio n of t he
re servo ir p erf o rm an ce ,
Influ x b ut o ne t h at is ve ry
u sef ul in p rac tice

Material Balance Equation
N B oi = (G - G pi ) B q - GB qi + G inj B qinj
[ NR si - (N - N p ) R s - G ps ] B q +
(N - N p ) Bo +
Schlumberger Public
NBoi
(C f + Cw Sw ) ( ) ( pi - p) +
1- Sw
W e - W p B w + W inj B winq
• ASSUMES:
– "ZERO DIMENSIONAL"
– NO HETEROGENEITY
– SINGLE AVERAGE PRESSURE
– NO CAPILLARY PRESSURE
– NO COMPOSITION CHANGE
Example
Schlumberger Public
Fluid Displacement
• 3. FLUID DISPLACEMENT :
Fractional Flow Concept
Schlumberger Public
Buckley-Leverett Equation
Welqe Tangent Method

Schlumberger Public

• Objective
To build a model of the reservoir and to examine its
performance in terms of production and pressure
Schlumberger Public
• Method
– The reservoir is divided into a number of blocks
– Basic data is provided for each block
– Wells are positioned within the arrangement of
blocks
– The required offtake rate is specified as a function
of time
– The appropriate equations are solved to give the
pressure and saturations for each block as well as
the production of each phase from each well
Pre-Simulation Era
Petroleum Reservoir
Performance Evaluation
via
Geo-
Schlumberger Public
physicist
Interaction of people
Integration of expertise
Reserv.
Eng.
Surface
Log Production Facility
Geologist Analyst Analyst Design

Post-Simulation Era
Surface
Facility Production
Production Geo- Analyst
Design
Analyst physicist
Log
Analyst
Schlumberger Public
Reserv.
Eng. Reserv.
Geologist Eng.
Geologist
Log Surface
Geo- Analyst Facility
physicist Design

Reservoir Simulation and Reservoir
Management
History
Match
N
Up-
Characterisation
scaling
Reservoir
Schlumberger Public
Reservoir Surface Facility Technically
Simulation Design/Update Feasible
Y
Well
Productivities
Economically Economics
Economics
N Feasible
Operations
Y
Preliminary 2M
NUMERICALPRODUCTS Workflow2H
Scaled-up
Simulation
Models 2N
STATIC Properties,
Grid Model Initial Model
2A
MODEL
Resources
2I Calibrated
Constraints
Models 2O Volumetric
2D Reservoir
Model NUMERICAL
Compart
Pressure
Production ments FORECASTING
Transient
Mechanism MODEL
Models
Schlumberger Public
Models Aquifer 2J
2B 2E Model 2P Revised /
Material
History
Balance Matched
Model SCAL Model 2K Model
(Kr,Pc, Comp)
Surface Network
2F
Constraints, Well
PVT Constraints
2R
Model Well 2Q Field
Production and Environment
2L / Injection Development
Pressure 2G Constraints
Data Base Model
Model
Reliability YES
ANALYTICAL
For
2C Forecasting FORECASTING
MODEL
Statistical and Analytical
Evaluation
March 06 NO Applied Reseervoir Simulation Day 1 40
PRODUCTS - Revised / History Matched
Model Revised
Reservoir Flow
Model Adjustments
Model To Reservoir
Calibrated Representation Properties
Reservoir Of Historical
Flow Model Performance
NEED
IMPROVEMENT
Comparison
Schlumberger Public
Model Vs
NO
Reported
Model
-Rate Schedules History Use History
SimOpt MATCH
-Well Completions Matching ?
Matched
-Re-completions Reservoir
-Layer Allocations Model
YES Flow Model
MATCH
NEED
Redefined SimOpt IMPROVEMENT
Model with Parameters Final
SimOpt Comparison
Parameters Model Vs
Well History NO
Reported
-Deviation
-Tubulars
Results from YES
-Completions Hessian, Co-
-Workovers variance and
Results from Sensitivity
Regressed
Analysis
Gradient Analysis Complete? Model
March 06 Simulation
Applied Reseervoir Simulation Day 1 41
Forecasting Model MECHANISM
IMPLEMENTATION
PLANS
FROM ROCK MODEL
SCAL FUNCTIONS FUNCTIONS
ADJUSTMENT -KR -KR
PROCESS -PC -PC
MODEL ORIGINAL TRANSPORT
CURRENT OR
COMPARISON OR MARKET
-SATURATION DIST PLANNED
MODEL VS CONSTRAINTS
-PRESSURE DIST. PIPELINE CPY.
FROM REPORTED
GRID SENSITIVITY -FLUIDS IN PLACE
PRELIMINARY -RADIAL/SECTOR
SIMULATION -CROSS SECTION
PROCESS
NEED
IMPROVEMENT MATCH
CURRENT OR SURFACE
PLANNED FACILITY
MODEL PLANT CPY. CONSTRAINTS
ADJUSTMENTS -RATES
GRID DESIGN TO RESERVOIR -PRESSURES
PROPERTIES
Schlumberger Public
GEOLOGIC MODEL
FROM -RESERVOIR GEOMETRY
SHARED EARTH -FAULT PLANE POSITION
MODEL HISTORY
-MATRIX PROPERTIES REVISED
PROCESS INITIAL MATCHED
-FRACTURE PROPERTIES RESERVOIR FORECASTING
RESERVOIR RESERVOIR
-WELL LOCATIONS FLOW MODEL MODEL (S)
FLOW MODEL FLOW MODEL
PROPOSED
RESERVOIR MODEL FORMULATION
MECHANISMS
ADJUSTMENTS
TO RESERVOIR
PROPERTIES
FLUID PROPERTIES
VERSUS PRESSURE OR NEED WORKOVER
FROM EOS DESCRIPTION IMPROVEMENT MATCH OR DRILLING
FLUID -WATER PROPERTIES CONSTRAINTS
PROPERTIES -ROCK COMPRESSIBILITY -FREQUENCY
MODEL
PROCESS -INJECTANT PROPERTIES -TECHNIQUE
REPRESENTATION
MODEL COMPARISON -RIG AVAIL
OF HISTORICAL
-RATE SCHEDULES PERFORMANCE MODEL VS
-WELL COMPLETIONS REPORTED
WELL TEST DATA -RECOMPLETIONS
INITIAL CONDITIONS RESERVOIR
-DST -LAYER ALLOCATIONS
-CONTACT POSITIONS SUMMATION
-PRODUCTION
-DATUM PRESSURE OF EXISTING
-PRESSURE
AND INFILL
MEASURED
BHP / THP AND
WELL HISTORY WELL
RATES DATA
AQUIFER -DEVIATION FLOW STRING CONSTRAINTS
DESCRIPTION -TUBULARS PERFORMANCE -PI OR II
-COMPLETIONS FUNCTIONS -BHP / THP
-WORKOVERS -MAX. GAS
HISTORY DATA
-PRODUCTION -MAX. WATER
PROD DECLINES
-INJECTION WELL CAPACITY
March 06
-PRESSURE
PRESSURE TRANSIENT TEST DATA
Applied Reseervoir Simulation Day 1
PROD / INJ ‘TIVITY
DAMAGE / STIMUL
42 -MDA OR PASS
-WELL DECLINES
Overview of the Reservoir
Schlumberger Public
Simulation Workflow and ECLIPSE
Software

Workflow
• The key to learning reservoir simulation is
to understand the workflow involved in
making a simulation study.
Schlumberger Public
• We will see various pieces of this process
during the course – this is an overview.
• The following will show the workflow and

software involved.

At the Beginning
• What do we have at the start of this process?

– Static model
• Structure – tops, thickness, layering, faults, boundaries, shales,
sands, rock type, depositional environment, grain size
Schlumberger Public
distribution fractures (all locations), property of aquifers
• Rock properties at all points in the reservoir (know at the wells,
estimated between wells) – permeabilities in all directions,
porosities, capillary pressure, relative permeabilities
• Initial Reservoir Conditions – water, oil, and gas saturations,
pressures at all points in the reservoir, contacts

At the Beginning
• What do we have at the start of this
process?
– Well Data
Schlumberger Public
• Well locations, trajectory, completions,
workover schedule
• Production rates of oil, water, and gas as a
function of time
• Pressure history of the wells – bottom hole
flowing pressure and/or build-up pressures
(well tests) at specific times
• Injection history – rates, fluids, pressures, etc.

At the Beginning
• What do we have at the start of this

process?
– Fluid data – PVT experiments with the
Schlumberger Public
reservoir fluids, viscosities, densities
– Material Balance of the reservoir history
– Reservoir Compartments
– Reservoir mechanisms
– Surface facilities and conditions

Workflow
Well
Production
data Grid building History Matching
FloGrid /GridSim(GRID) SimOpt
FINDER
Data Processing
Schlumberger Public
OFM PVT-data
PVTi Simulation Analysis
Well data
preparation Office
Schedule ECLIPSE /
FloViz
Rock data
FrontSim (Graf)
SCAL
Wellbore hydraulics
VFPi Well Planning
Geological PlanOpt
model Well Testing
Weltest
GEOFRAME Near Wellbore modelling
CPS3 NWM
Petrel

First Step in the Workflow
• FloGrid: From the Static Model-

Create a numerical grid of the
reservoir (flow field) that includes at
Schlumberger Public
all points
– Rock permeability, porosity, relative
permeability, capillary pressure,
pressure saturations
– Location, volumes, adjacent rocks or
features

Schlumberger Public
Next Step in the Workflow
• PVTi: PVT Analysis and Fluid

Characterization
– Using Equation of State and PVT
Schlumberger Public
Program characterize oil and/or gas and
match laboratory phase behavior
experiments
– Export PVT files for Blackoil,
Compositional, or Thermal simulations

Schlumberger Public
Next Step in Workflow
• SCAL: Analysis of Rock Types,

Relative Permeabilities, Capillary
Pressure, Imbibition and Drainage
Schlumberger Public
and Assignment of curves to grid
blocks

Krow D
Krw I
Krow I
Krw D
Schlumberger Public
Krog
Krg Imbibition
Krg
Imbibition Drainage

• Schedule: Well Data Preparation

– Location of wells in the grid
– Averaging and Assignment of
Schlumberger Public
production rates to wells
– Workover wells
– Assign all well, control, economic limits
and time stepping keywords to ECLIPSE
data set

Schlumberger Public
• VFPi: Create Table relating BHP to

THP as a function of
• Well design – depth, ID, roughness, depth,
Schlumberger Public
temperatures,
• GOR, Water Cut
• Oil rates
• Pumps, gas lift, compressors
– Simulator need BHP – if you want to
control well from the surface must have
lift curves

Schlumberger Public
Now we have all the initial data
Schlumberger Public
ready for an ECLIPSE simulation

• ECLIPSE Office: Putting everything

together – controlling version –
controlling runs
Schlumberger Public
Schlumberger Public
Office Run Control
Schlumberger Public
• Simulations –
– Blackoil
– Compositional
Schlumberger Public
– Thermal
– Special
• Solvent – Miscible
• Dual Porosity
• Polymer
• ECLIPSE Advanced Options – see next
slides

Advanced ECLIPSE Options
• API Tracking
• Coal Bed Methane
• Environmental Tracers
Schlumberger Public
• Flux Boundary Conditions
• Foam Model
• Gas Field Operations
• Gas Lift Optimization
• Geomechanics

• Hysteresis
• Local Grid Refinement and
Coarsening
Schlumberger Public
• Miscible Flood Modeling
• Multi-Segment Wells
• Network Option
• Nine-point Schemes
• Parallel Option
• Polymer Flood Model

• Reservoir Coupling
• Solvent Model
• Surfactant Model
Schlumberger Public
• Vertical Equilibrium
• Thermal Option
• Wellbore Friction Option
• Pseudofunctions

• SimOpt: Computer Aided History

Matching Using SimOpt
– Calculates gradients
Schlumberger Public
– Regression/Optimizer changes HM
parameters to minimize HM Objective
function

Schlumberger Public
• FloViz: View Reservoir properties

and solutions as a function of time
(saturations, pressures, Rs, Rv, mole
Schlumberger Public
fractions,….) in 3-D.

Schlumberger Public
After History Matching – Future
Schlumberger Public
Operations are Simulated

ECLIPSE Workflow Finished
• Economics – Peep Workflow next

step
Schlumberger Public
Schlumberger Public
End of Workflow Overview

Our Virtual Reality
∂  kkrp ∂Φ  ∂  kkrp ∂Φ p 
+ +
   
∂x  µ p B p ∂x  ∂y  µ p B p ∂y 
∂  kkrp ∂Φ p  ∂  φSp 
=
∂z  µ p B p ∂z  ∂t  B p 
 
Schlumberger Public
Grid Block
Reservoir Parameters (~100mx100mx10m)
to be assigned
Permeability, k Porosity, φ Thickness, h

Elevation, d Saturation, S Pressure, p
Compressibility, c Fluid PVT RelPerms, kr
Well data, q, p

Which Simulator?
IMPES Isothermal
Temperature
Linearization
IMPLICIT Thermal-BOS
Schlumberger Public
AIM Thermal-Comp
Classification
Simulators
of
Co-ordinates Systems
Cartesian-orth
Fluid Description
Black Oil Radial-orth

EOS-Comp Corner Point
Chemical
Un-Structured

Where ECLIPSE Fits In
Static View of
the Reservoir Dynamic Field Data
from from
THE LINK
Schlumberger Public
Geophysics Production Wells
and IS Pressures
Geology Oil Production Rates
ECLIPSE
and Water Cuts
Petrophysics GOR

What is ECLIPSE?
Fully
Fully
Implicit
Implicit
Stateof
State ofthe
the
Art
Art
Simulator
Simulator
Schlumberger Public
Stable
Stable
Robust
Robust
Accurate
Accurate
Flexible
Flexible
Dominates
Dominates
>85%
>85%
ofofthe
the
Market
Market
What Modeling can ECLIPSE do?
Hydrocarbon Complex Multi-phase
Thermodyn. Geometries Flow
Thermal
Schlumberger Public
Multiple
Contact ECLIPSE Processes
Miscibility State of the
art
Simulator Tracer
Vertical Movement
Equilibrium
Polymer
Fractured Surfactant
Reservoirs Gas Field Coalbed
Operations Methane
Overview of Modeling Procedure
Property 3D, FloGrid Describe reservoir structure (seismic, logs)

CPS3, Petrel reservoir gross and net thickness (logs)
well locations and perforated intervals
VFPi
FloGrid
Schlumberger Public
Design reservoir porosity, permeability (logs, cores)
PVTi grid fluid analyses (lab data)
pressure and contacts (logs, well tests, etc.)
Schedule
SCAL
Select simulator black oil or compositional
ECLIPSE fractured, condensate, etc.
model
100/300/500 horizontal wells, EOR, thermal, etc.
Solve for
ECLIPSE History
pressures and historical production data
match
SimOpt saturations
Predict and investigate different scenarios

FloViz optimize future visualize results
production economic calculations
Peep

“Fluid Flow Simulation” and
“Material Balance”
• A sad story.....
• Many engineers working on simulation think they

are doing -
Schlumberger Public
⇒ Fluid Flow Simulation
• BUT
• they are actually doing -
⇒ Material Balance
• ... even then they are using a reservoir simulator

What Benefits?
• Golden Rule
– You can only produce once
– You can simulate many times
• Relative merits of different recovery mechanisms
Schlumberger Public
• Maximizing hc recovery
• Type of water flooding
• Well location and spacing
• Production versus hydrocarbon recovery
• Gas deliverability
• Best completion schemes for wells
• Section of the reservoir from which oil is produced

Conducting a Study
Building
Simulation Model Reporting
Predicting
10% 5%
10%
Schlumberger Public
Building Reservoir
Model
30%
Matching History
Define Problem 40%
5%

Schlumberger Public
Characterizing the Reservoir

Building Simulation Model
Reservoir Characterization
Top
Surface
Map
Schlumberger Public
Rock Volume
Structural
Isopach
Map
Map

Data Structures for Static Model
• Data for the Static Model can come
in 3 different types
Schlumberger Public
1. Deterministic ( Fluid contacts?
PVT? Areally extensive rocks? )
2. Stochastic with trend ( Porosity

permeability relationships?)
3. Stochastic with no trend

Data Sources: What Data & Where
to Get it?
POROSITY CORE, LOG AND WELL
TESTING
PERMEABILITY CORE, LOG, CORE
CORRELATED LOG AND
WELL TESTING
RELATIVE SPECIAL CORE ANALYSIS,
PERMEABILITY AND WELL TESTING
Schlumberger Public
CAPILLARY PRESSURE CORE DATA, LOG DATA
SATURATION SINGLE WELL TRANSIENT
TESTING, LOGGING,
SPECIAL CORE ANALYSIS
GRAIN SIZE THIN SECTIONS, NUCLEAR
DISTRIBUTION MAGNETIC RESONANCE, X-
RAY, CAT
PORE CORE ANALYSIS
COMPRESSIBILITY
CATION EXCHANGE CORE ANALYSIS, WELL
CAPACITY LOGGING
CLAY TYPE AND WELL LOGGING, CORING
CONTENT
PRESSURE RFT AND DST
RESERVOIR SEISMIC, WELL LOGS
GEOMETRY

Data Scales
Method Typical Relative
Volume Volume
Scale (cft) Reservoir=1
Geologic Thin 3.5x10-6 8x10-17
Section
Core Plug 0.000682 15.5x10-15
Core 0.0873 0.2x10-11
Schlumberger Public
( 4” diameter
and 1 ft length )
Wireline Log 7 16x10-11
(neutron
porosity, density
and gamma ray)
for two vertical
feet
Well test 0.44x109 .01
( radius = 1183,
h=100 ft)
• Most of our data looks at a very small piece of

the reservoir. Applied Reseervoir Simulation Day 1
March 06 87
Schlumberger Public
Benefits of Reservoir Simulation

Reservoir Simulation Benefits
• Examine the performance of a given

reservoir under depletion, water
injection or gas cycling.
Schlumberger Public
• Judge the most effective type of
water flooding: relative merits of
– flank water injection
– pattern water flooding.

• Estimate the effect of well location

and spacing and number of wells.
Schlumberger Public
• Estimate the effect of production rate
on the hydrocarbon recovery and
economics.

• Maximize the hydrocarbon recovery.
• Find optimal methods of field
Schlumberger Public
development and production
schemes.
• Assess possible enhanced oil

recovery (EOR) schemes and their
implementation.
• Assess laboratory and field data

requirements and their effect on the
performance predictions.
Schlumberger Public
• Investigate the reason why the
reservoir behavior deviates from the
earlier predictions.

• Establish the best completion

schemes for wells.
Schlumberger Public
• Identify the section of reservoir from
which the hydrocarbon is produced.

Typical Problems
– How Many Wells
– What Rate
– Infill Drilling
Schlumberger Public
– Perforation
– Work-over
– Pressure Maintaince
– Water or Gas Injection
– Pattern Flood
– Peripheral Flood

Objectives of Simulation Study
1. To gather all information available
2. To augment/replace basic methods
3. To history match past performance
4. To predict future performance
Schlumberger Public
5. To determine what is important
6. To examine effects of variation in data
7. To compare scenarios
8. To investigate problem areas
9. To understand the reservoir
requirements for simulation

Requirements for Simulation
• 1. Knowledge of the Reservoir
• 2. Statement of the Objectives
• 3. Critical examination of the Data
• Cautions :
Schlumberger Public
1. Understanding of the reservoir is essential to building a model
2. The model should be made as simple as possible to answer
the vital questions
3. The amount of output that can easily be absorbed must be
considered
Rubbish in gives rubbish out
Garbage in give garbage out

Data Required at Each System
Node
1. Permeability
2. Porosity
3. Thickness
Schlumberger Public
4. Elevation
5. Initial Saturation for Each Phase
6. Initial Pressure
7. Fluid Properties (oil, water, gas) B µ Rs Rv
8. Rock Properties Kr vs. S, Pcow Pcgo Cf
9. Grid Dimensions

Other Required Data
1. Well Producing Interval and Productivity
2. Pressure Loss in Tubing and Flow Lines
Schlumberger Public
3. Aquifer Description
4. Historical Oil, Water, and Gas

Production
5. Observed Pressure versus Time

Schlumberger Public
Formulations of Equations

Terminology in Equations and
Solution of Equations
Schlumberger Public
Overview of the discretization
Q+z
process Q-y
3-D Flow
Directions
Q-x Q+x
Q-y
Q-z
Schlumberger Public
Identifying
Blocks
i i+1
Transmissibility
 kA   k r  i+1/2
   
 ∆ 1   µβ  Block
Non-neighbor
Interface
Connections
Controls interblock flow
I.e. Faults
Stirred Tank – Diffusive Process in
Schlumberger Public
Continuum Concept and Scale
Pore space
1
C o n tin u u m
D o m a in
Schlumberger Public
φ
R e p re s e n ta tiv e
E le m e n ta ry
V o lu m e
0
V o lu m e
Sand grain

Schlumberger Public
Review of Mathematics – 1
• Differential operators
• In 3 dimensions velocity is a vector
G G G G
Schlumberger Public
u = uxi + uyj + uzk
• If P is a scalar function, then ∇P

(gradient of P) is a vector
∂P G ∂P G ∂P G
∇P = i + j + k
∂x ∂y ∂z

Review of Mathematics - 2
example 1:
let P = 5 x that is pressure increases
as 5 times the x coordin ate
Schlumberger Public
∇P = i + j+ k
∂x ∂y ∂z
∂P G ∂ G G
∇P = i = (5 x )i = 5 i
∂x ∂x
G
the gradient of pressure in the x (or i ) direction is 5

example 2:
let P = 5 x + 3 y + 1z that is pressure increases
as 5 times the x cooridi nate, 3 times the y coord.. ..
Schlumberger Public
∇P = i+ j+ k
∂x ∂y ∂z
∂P G ∂P G ∂P G ∂ G ∂ G ∂ G G G G
∇P = i+ j+ k = (5 x )i + (3 y ) j + (1z )k = 5i + 3 j + k
∂x ∂y ∂z ∂x ∂y ∂z
G
the gradie nt of pres sure in t he x (or i ) directio n is 5
G
the gradie nt of pres sure in t he y (or j ) directio n is 3
G
the gradie nt of pres sure in t he z (or k ) directio n is 1

• If u is a vector, then ∇· u (divergence

of u) is a scalar
Schlumberger Public
G ∂ux ∂uy ∂uz
∇ ⋅u = + +
∂x ∂y ∂z

example :
G G G G
velocity u = 5 x i + 3 yj + 1zk
thus, u x = 5 x , u y = 3 y, uz = z
Schlumberger Public
G ∂ux ∂uy ∂uz
∇ ⋅u = + +
∂x ∂y ∂z
∂ ∂ ∂
= (5 x ) + (3 y ) + (z )
∂x ∂y ∂z
G
= 5 + 3 +1 = 9 t he divergence of u is 9

• Special case: If P and f are scalars,

then the divergence of f times the
gradient of P is a scalar
Schlumberger Public
∂  ∂P  ∂  ∂P  ∂  ∂P 
∇ ⋅ ( f∇ P ) =  f  +  f  +  f 
∂x  ∂x  ∂y  ∂y  ∂z  ∂z 

• For the special case where f = 1, we

get the Laplacian of P
∂ ∂ ∂
Schlumberger Public
2 2 2
P P P
∇
2
P = 2 + 2 + 2
∂ x ∂ y ∂ z
• Note for incompressible flow, single

phase, no wells (sources or sinks)
∇ Φ = P − ρgz
2
Φ = 0 where
∂  k x ∂Φ  ∂  k y ∂Φ 
V  φSo 
 + n +1 n
∂x  µβ ∂x  ∂y  µβ ∂y 
a a
 φS  
PDE F=
∂f F(x + 2 ) - F(x - 2 )
; ∆[To ∆Φ o ] =   -  o  
∂  k ∂Φ  ∂ (φ/β )
+  z
∂x a a ∆t  Bo   Bo  
=
∂z  µβ ∂z  ∂t
(x + ) - (x - )
2 2 
Finite difference
equation
Linearise Discretise Solve the

Derive
Schlumberger Public
the the Linear
Equations
Equations Equations Equations
a11x1 + a12 x 2 + .......... + a1n x n - b1 = 0

(AB)k +1 = (AB)k + A k +1∂B + Bk ∂A
a 21x1 + a 22 x 2 + .......... + a 2n x n - b 2 = 0
.
.
.
a n1x1 + a n2 x 2 + .......... + a nn x n - b n = 0

The Continuity
Equation
Schlumberger Public
Constitutive The Equation
The
Equations of Flow
Equation
of State

Simulation – The key is to determine
the inter block flow – given the cell
properties
Producer
inter block flow q= A k kr ∆P
∆x µ
Schlumberger Public
Cell Properties:
Structure - Depth, Thickness, DX, DY
Properties - Permeability, Porosity

Conservation of Mass, Single
Phase
Control Volume
Cross-sectional Area = A
Flow x
Schlumberger Public
∆x
• Volume of Control Volume (CV) = A * ∆x (m3)

• Rock porosity in Control Volume = φ (t)
• Fluid density (mass/volume) = ρ (x,t)

Conservation of Mass, Single
Phase
• Mass of fluid in Control Volume

• = m = ρ • φ • A ∆x = ρ • φ • ∆V
Schlumberger Public
• Mx = Mass flux in or out of CV= mass
flow/area/time
• Units: (kgm/ m2 · sec)
∆t = time interval

Mass Balance Equation
• Over the time interval ∆t the Material
Balance Equation is
• IN - OUT = ACCUMULATION ± injection/production
Schlumberger Public
• IN = Mxx • Α • ∆t
• OUT = Mxx+∆ x • Α • ∆t
• Accumulation = mass in placet+∆ t - mass in placet
• = [ρφ • A • ∆x]t+∆ t - [ρφ • A • ∆x]t
• Mass accumulation is due to compressibility as

the pressure changes.
Fluid compressibility is represented by ρ = ρ(P)
Rock compressibility is represented by φ = φ(P)
Inj/Prod = Q • Α∆x • ∆t
Schlumberger Public
Thus: IN - OUT = ACC ± Inj/Prod becomes
[Mx x - Mx x+∆ x ] • A ∆t = [ρφ • A • ∆x]t+∆ t - [ρφ • A • ∆x]t
± Q • Α∆x • ∆t (EQU 1)

Divide EQU 1 by: Α∆x • ∆t we get
Mx x - Mx x+∆ x = ρφ t+∆ t - ρφ t
+Q EQU 2
∆x ∆t
Schlumberger Public
Take Limit as ∆x 0
Take Limit as ∆t 0
As per Calculus

• The result is
∂ Mx ∂
− = ( φρ ) + Q EQU 3
∂x ∂t
Schlumberger Public
• We get partial derivatives because
• Mx = f(x,t) and φρ = f(x,t)

Momentum Balance
• Next we need a relationship between
velocity and pressure
• In porous media we use Darcy’s Law (see
Schlumberger Public
next slide)
• ux = Darcy Velocity in the x-direction
Mass Flux = Mx = ρux (EQU 4)
(Units: kgm/m3 • m/sec = kgm/ m2 • sec)

Schlumberger Public
Henry DARCY
Schlumberger Public
Henry Philibert Gaspard Darcy lived from 1803 to 1858 and
was a well respected and renowned scientist of his time. His
name is still known among hydrologists. He is best known for
his contribution to groundwater flow: the Darcy "Law" that
groundwater flow is directly proportional to the pressure
difference. However, he made substantial contributions to a
much broader field of hydrological and hydraulic engineering.
Schlumberger Public
As an engineer responsible for the water supply of Dijon he
carried out experiments to determine the characteristics of
water flow through pipes of different material, size and
roughness; and did filtration tests through sand filters of
different sizes and filter material. He also carried out field
experiments in open channels to determine the relations
between velocity, cross-sectional area and slope. As a result of
these experiments his name also appears in the Darcy-
Weisbach equation for open channel flow.
Table of the experiments made in Dijon October 29 and 30, and November 2, 1855.
Experiment Duration Mean Flow Mean Ratio of OBSERVATIONS

Number min l/min Pressure volume
m s and
pressur
e
1st Series, with a thickness of sand of 0.58 m
1 25 3.60 1.11 3.25 Sand was not washed
Schlumberger Public
2 20 7.65 2.36 3.24 
3 15 12.00 4.00 3.00  The manometer column
4 18 14.28 4.90 2.91  had weak movements
5 17 15.20 5.02 3.03 
6 17 21.80 7.63 2.86 
7 11 23.41 8.13 2.88  Very strong oscillations.
8 15 24.50 8.58 2.85 
9 13 27.80 9.86 2.82 Strong manometer
10 10 29.40 10.89 2.70 oscillations.
2nd Series, with a thickness of sand of 1.14 m
1 30 2.66 2.60 1.01 Sand not washed.

2 21 4.28 4.70 0.91
3 26 6.26 7.71 0.81
4 18 8.60 10.34 0.83 
5 10 8.90 10.75 0.83  Very strong oscillations.
6 24 10.40 12.34 0.84 

Schlumberger Public
Schlumberger Public
Permeability Units
• In Russia, for example, permeability

is give in units of Meters2 or cm2
Schlumberger Public
1 Darcy = 1,000 mD = 0.9869x10-8 cm2

Permeability Tensor
Y kv
k
Permeability is second rank symmetric
Schlumberger Public
tensor
k xx k xy k xz 
 
[k] = k yx k yy k yz 
U Angle 
k k k

 zx zy zz 
P1
where kxy = kyx, kxz = kzx and kyz = kzy.
P2
x
1 Cos 2α Cos 2 β Cos 2γ

= + +
k kx ky kz

Darcy’s Law – Single Phase Flow
• 1-D Horizontal Flow
k ∂P
u = − ⋅ EQU 5
Schlumberger Public
x
µ ∂x
• 3-D - 1 Phase Flow with Gravity
G
G k  g 
u = − ∇P + ρ ⋅  EQU 6
µ  gc
• where gc is a conversion constant, note z
is positive down

Darcy’s Law
• We know that
G
g g
Schlumberger Public
ρ = ρ ∇z = γ∇z EQU 7
gc gc
• Thus Darcy‘s Law in 3-D vector form:
G k
u = − ( ∇ P − γ∇ z ) EQU 8
µ
Darcy’s Equation and Flow in
Generalized Darcy’s Law Porous Media
k xx dΦ k xy dΦ k xz dΦ
Vx = - - -
µ dx µ dy µ dz Simplification
k yx dΦ k yy dΦ k yz dΦ
Vy = - - -
µ dx µ dy µ dz If the principal axes of
Schlumberger Public
permeability coincides with
k zx dΦ k zy dΦ k zz dΦ
Vz = - - - the axes of the co-ordinates
µ dx µ dy µ dz
system, the cross terms
or disappear.
 Vx   k xx k xy k xz   J x  k xx 0 0 
      
 Vy  =  k yx k yy k yz   J y   0 k yy 0 
 V  k  
 z   zx zy zz   J z 
k k 0 0 k 
 zz 
- 1∂ φ
J= , 1 = x, y, z
where µ∂1
Combining Darcy’s Law with
Material Balance
• We substitute ux for Mx in EQU 3 and
get the following
∂ ( ρux ) ∂
Schlumberger Public
− = ( ρφ ) + Q
EQU 9
∂x ∂t
• In 3 dimensions
∂
G
− ∇ • (ρu ) =
∂t
(φρ)+ Q EQU 10

Combining Darcy’s Law with
Material Balance
• We substitute for ux (from EQU 8) for
the 3-D with Gravity
Schlumberger Public
 ρk  ∂
∇ ⋅  ( ∇P − γ∇z )  = ( ρφ ) + Q
µ  ∂t
EQU 11

Simulator Flow Equation
• Now if we divide EQU 11 by ρSTC and

define the formation volume factor B
= VRC/VSTC we get the following:
Schlumberger Public
∂ φ 
[
∇ ⋅ λ (∇ P − γ∇ z ) ] =
∂t  B  + q
EQU 12
k  Q
• Where λ = and q =  ρ 
µB

Multiphase Simulation Equation
• Volatile oil: gas dissolved in oil - Rs,

• no oil vaporized in gas
Schlumberger Public
Stock Tank Oil: oil Reservoir Oil: oil components +
component dissolved gas (Rs)
Stock Tank Gas: gas Reservoir Gas: gas component
component
Reservoir Water: water
Stock Tank Water: water component
component

Black Oil model Phase Options
Phase Combination RUNSPEC Keywords
Live oil with OIL, GAS, DISGAS,
3 Water
dissolved gas WATER
Wet gas with GAS, OIL, VAPOIL,
3 Water
Schlumberger Public
vaporised oil WATER
Live oil with Wet gas with OIL, GAS, DISGAS,
3 Water
dissolved gas vaporised oil VAPOIL, WATER
2 Dead oil Water OIL, WATER
2 Water Dry gas GAS,WATER
2 Dead oil Dry gas OIL, GAS
1 Dead oil OIL
1 Water WATER
1 Dry gas GAS

Oil Formation Volume Factor / Rs
[Vo + Vdg ]RC Bo

Bo =
[Vo] STC
Schlumberger Public
P
[Vdg ]STC Rs
Rs =
[Vo] STC
P
Water and Gas FVF
[Vw] RC
Bw = Bw
[Vw] STC
Schlumberger Public
P
[Vg ]RC Bg
Bg =
[Vg ]STC
P

Schlumberger Public
Black Oil Equation-of-State
(Generalized)
ρo,sc + Rsρg,sc
Oil: ρo =
Bo
Schlumberger Public
ρw,sc
Water: ρw =
Bw
ρg,sc + Rv ρo,sc
Gas: ρg =
Bg

Simulation Equations – Multi-
Phase Flow
• For multi-phase flow of phase i
• 1. Continuity Equation
Schlumberger Public
∂ ∂
− ( ρ i u i )= (φ S i ρ i ) + Q i
∂x ∂t
• 2. Fluid Flow Equation ui = velocity

of phase i
kk ri g
ui = −
µi
( ∇Pi − γi∇z ) whereγi = ρ i
gc
Phase Flow
• Three phase (o-w-g) flow equations
• volatile oil: gas dissolved in oil - Rs,
• no oil dissolved in gas
Schlumberger Public
• Water component:
[
∇ ⋅ λ w (∇ P w − γ w ∇ z ) = ]
∂  Sw 
φ
 B w  + q w
∂t
Phase Flow
• Oil component:
[
∇ ⋅ λ o ( ∇ P o − γ o∇ z ) = ]
Schlumberger Public
∂  So 
∂t φ B o  + q o

Phase Flow
• Gas component:
[ ]
(1)
∇⋅ Rsλo( ∇Po − γo∇z) + λg( ∇Pg − γg∇z)

(2) (3) (4)
Schlumberger Public
(5) (6)
∂   So Sg   (7) (8)
= φ Rs +   + Rsqo + qg
∂t   Bo Bg  

Phase Flow
• where phase transmissibility, phase i
k ⋅ kri
Schlumberger Public
λi =
µi ⋅ B i

Phase Flow
• We have 3 component balances and
six dependent variables which are
the unknowns:
Schlumberger Public
• Po So Pw Sw Pg Sg
• Need 3 additional relationships:
• So + S w + Sg = 1
• Pcow = Po - Pw = f(Sw)
• Pcog = Pg - Po = f(Sg)
Unknowns Dependence on Phases
Present
There are 3 cases:
IF Sg > 0 and So > 0 this means saturated oil and gas

therefore Rs = R (Pg). Therefore we solve for Po, Sw, Sg.
Schlumberger Public
IF Sg = 0 this means that Rv = R (Pg) is not a required
and we solve for Po, Sw, and Rs. Rs vary from 0 to saturated.
IF So = 0 this means that Rs = R (Po) is not a required

and we solve for Po, Sw, and Rv. Rv vary from 0 to saturated.

Discretization
– Converts continuous PDE into difference form
– Replaces original problem with other problem,
which can be solved easily
– The reservoir domain is presented by spatially
Schlumberger Public
distributed, interconnected discrete elements
(grid blocks)
– Temporal (time) domain is also discretized
– The reservoir parameters are calculated over
these constitutive elements at discrete time
steps

Finite Differences
∂ ∂
Replace and
∂x ∂t
with point values at specific points in space and
Schlumberger Public
time.
An ``Engineering'' Approximation: Taylor Series:
∆x ∂f( x o ) ∆ x2 ∂2 f( x o )
f ( x o + ∆x) = f ( x o ) + +
1! ∂x 2! ∂ x2
∆ x3 ∂3 f( x o )
+ 3 + ...
3! ∂x

Finite Difference Approximations
to First Derivatives
forward difference ∂ f = f ( x + ∆ x ) − f ( x )
∂x ∆x
Schlumberger Public
central difference ∂f f (x + ∆x) − f (x − ∆x)
=
∂x 2∆x
Error terms:
∆x ∂ 2 f ∆x2 ∂ 3 f
forward error − 2 − 3 − .....
2! ∂x 3! ∂ x
∆ x 2 ∂ 3 f
central error − − .....
3! ∂x 3
Approximation to Second Order
Derivatives
∂  ∂f 
 =? Let x+a be forward direction
∂x  ∂x 
Schlumberger Public
Let x-a be backward direction
Let x+a/2 be forward direction in half a
Let x-a/2 be backward direction in half a
a a
F(x + ) - F(x - )
2 2 = f(x + a) - 2f(x) + f(x - a)
a a a 2
(x + ) - (x - )
2 2
Second Derivative
k ro
let M o = Then the second derivative
µ o Bo
viscous term becomes
Schlumberger Public
∂  ∂po 
 kx M o ≈
∂x  ∂x 
(k x M o )i +1/ 2, j [( po )i +1, j − ( po )i , j ] (k x M o )i −1/ 2, j [( po )i , j − ( po )i −1, j ]

−
xi +1 − xi xi − xi −1
xi +1/ 2 − xi −1/ 2

Transmissibility
• Flow from one grid block is

controlled by
– The pressure difference (pressure
Schlumberger Public
gradient) between grid blocks
– And the transmissibility (ease of fluid
flow) at the boundary between the grid
blocks.

Transmissibility
• Transmissibility has 2 components

– Fixed component function of geometry
– Variable component function of flowing
Schlumberger Public
phase saturation in the UPSTREAM Grid
Block

Fixed Component of
Transmissibility
flow
i-1 i
Schlumberger Public
TRANXi is transmissibility at this boundary
Dxi and Dxi-1 = the length of grid blocks
Dyi, Dyi-1, Dzi, Dzi-1 thickness, height of grid blocks
PERMXi and PERMXI-1 = x-permeability of each
block
A = Interface area between cell i an i-1
NTGi and NTGi-1 = Net To Gross in each block
Fixed Component of
Transmissibility
• Without dip in the reservoir
Dxi −1 ⋅ Dyi ⋅ Dzi ⋅ NTGi + Dxi ⋅ Dyi −1 ⋅ Dzi −1 ⋅ NTGi −1
Schlumberger Public
A=
Dxi + Dxi −1
Const ⋅ A
TRANX i =
1  Dxi Dxi −1 
 + 
2  PERMX i PERMX i −1 

Weighting of Variable Component
of Transmissibility
Flow
For the value

at grid block i
Schlumberger Public
• Downstream: use i and i - 1
• Mid-point: use i + 1 and i - 1
• Single point upstream: use i + 1 and i
• Two point upstream use i + 2 and i +1
and extrapolate saturation to i + 1/2
boundary
Influence of Mobility Weighting
Schlumberger Public
Finite Difference Simulation Equation
Oil-Water System
Oil Equation, X Dimension Only
 k x ∆ y ∆ z k ro 

 µ B ∆x 
 [ P n +1 o i - 1 - P n +1 o i + γ o ( z i - 1 - z i ) ]
 o o  i -1/2
 k ∆ y ∆ z k ro 
Schlumberger Public
-  x 
 [ P n +1 o i - P n +1 o i+ 1 + γ o ( z i - z i - 1 ) ]
 µ o Bo ∆ x  i+1/2
 φ   S on+1 + S on 
- q os δ i = ( ∆ x ∆ y ∆ z ) i    
 Bo  i  ∆t i
  So   d (1/ B o )    p n+1 - p n 
+( ∆x ∆y ∆z )i    (1 − φ )C f + ( φ S o ) i  
  
o o 

  Bo  i  dp i   ∆t i
NOTE: n = current time level, n+1 = new

(unknown) time level

Water Equation, X Dimension Only
 k x ∆ y ∆ z k rw 
 µ B ∆  [ P oi-1 - P oi - Pcowi -1 + Pcowi + γw ( z i -1 - z i )] -
  n +1 n +1
 w w x  i -1/2
 k x ∆ y ∆ z k rw 

 µ B ∆ 
 [ P n+1 oi - P n+1oi+1 + Pcowi - Pcowi+1 + γw ( zi - zi+1 ) ]
 w w x  i+1/2
Schlumberger Public
 φ   S on+1 - S on  This is unknown since it
- q ws δ i = ( ∆ x ∆ y ∆ z ) i    
 B w i  ∆t i is at the new time level
 (1 - S oi )  Poin+1 - Poin 
- ( ∆ x ∆ y ∆ z )i  (1 − φ )Cf  
 B wi  ∆t 
 d (1/ B w )i   Pon+1 - Pon Pcown+1 - Pcown  

+ φ i (1 - S oi )  
  -  
 dp   ∆ t ∆ t i 
NOTE: n = current time level, n+1 = new (unknown) time level

NOTE: 1 Equation, 2 Unknowns, Pon+1, Son+1
Linearization and Solution
Schlumberger Public
Process

Linearization and Solution
Process
• There are 2 types of linearization
that we must understand
• We will discuss both of these
Schlumberger Public
Linearization (1) in the PDE
 ∂S o 
in the PDE we have   which becomes
 ∂t 
 S o n +1 − S o n 
  in the finite difference equation
 
Schlumberger Public
 ∆t i
We would like to replace this unknown derivative with known

quantities, that is, at the nth time level
We use the chain rule and the fact that capillary pressure
varies slower than viscous pressure and is evaluated at the
OLD time level (n).

• To Linearize: Expand the time

derivatives
 ∂ S o   dS o   ∂ Pc 
  =    
Schlumberger Public
 ∂ t   dP c   ∂ t 
 dSo 
Where   is considered to be a unique function of So (of Pc ).
 dPc 
Now P C is evaluated explicitly , thus
n n
∂S n +1
 dS   ∂ Pc 
o
=  o
   (EQ 56)
∂t  dP c   ∂t 

• When we replace the term evaluated

at the n+1 time level with the product
of 2 terms evaluated at the nth time
Schlumberger Public
level we have linearized the PDE
• The only remaining unknown is the

pressure at the new time level that is,
Pn+1

Linearization Part 2
• We now have an equation with one

unknown BUT ------
• We find that several of the
Schlumberger Public
coefficients in front of the unknown
pressures are functions of the
pressure at the new time level
(depending on the level of
implicitness in the solution
procedure)

Linearization Part 2
• Thus, to solve this equation with

non-linear coefficients we must
“linearize” again.
Schlumberger Public
Linear and Non-Linear
Ax + By = D this is linear
where x and y the unknowns
and A and B are constants
Schlumberger Public
A( x, y ) x + B( x, y ) y = D this is non - linear
since A and B are functions of the unknow solutions x and y

Linearize (2) the Equation
Let A be the iteration counter

We know the solution at A, we are looking
for the solution at the next non - linear iteration A + 1
Schlumberger Public
We linearize the equation by evaluating the
coefficents A(x, y) and B(x, y)
at the A iteration level where x and y are known.
[A( x , y )] x
A A A A +1
[ ]
A
+ B( x , y ) y A +1 = D
A A

Convergence of the Non-Linear
Iterations
If this solution process converges
(x − x ) → 0 and
A +1 A
( y − y ) → 0 and
A +1 A
Schlumberger Public
[A ( x , y ) = A ( x
A +1 A +1 A
, yA) ]
[B ( x , y ) = B ( x
A +1 A +1 A
,y A
)]
• The coefficients do not change any more,

so the solution to the linearized equation
is the correct solution to the non-linear
equation.

• Have system of Non-Linear Finite
Difference Equations to solve at each grid
block
Schlumberger Public
• Unknowns P, Sw, Sg
• To solve the non-linear system:

– Linearize equations
– Solve the linear system for P or P, Sw,Sg
– Converge the non-linear system with Newton
Iterations

Matrix Inversion
• Matrix Inversion - to solve the
simulator equations efficiently
a11x1 + a12 x 2 + .......... + a1n x n - b1 = 0
a11 a12 . a1n   x1  b1 
Schlumberger Public
a 21x1 + a 22 x 2 + .......... + a 2n x n - b 2 = 0
a a 22 . a 2n  x  b 
.
 21  2  2
.
.
A = . ; x = . ; b = . 
     
a n1x1 + a n2 x 2 + .......... + a nn x n - b n = 0
.  .
  . 
a n1 a n2 . a nn  x n  b n 
    
• Simultaneous Linear Equations

Corresponding Matrix Equation
Linear Problem
AP = d
Schlumberger Public
a b c d   P1   d1 
e f g h   P2  d 2 
i j k l   P3  =  d 3 
 
m n o p   P4  d 4 

Iteration Process in Reservoir
Simulation
Schlumberger Public
Example of linear
and non-linear
iteration process:
4 non-linear
iterations Usually a non-linear iteration
requires 20 to 30 linears to
converge pressure and saturations

Linearization
• Newton-Raphson Technique: Y
• It is primarily a root finding technique
Schlumberger Public
Y
X
Root x4 x3 x2 x1
Newton-Raphson convergence course

X
x3 x1 Root x0
x2
f(x k )
∂f(x k )
x k +1 = x k -
Divergent Newton-Raphson iteration process
∂x

Solution to the Non-Linear
Problem
• Have choices:
– can solve unknowns one at a time using
parameters (k, kr, Pc, µ, ρ) at current (n,
Schlumberger Public
old) time --- called explicit
– can solve all unknowns simultaneously
using parameters at new (n+1, new)
time --- called implicit or fully implicit

Problem
• Explicit - has severe stability
constraints
– for example throughput for a grid block
Schlumberger Public
can not exceed 5 to 10% of the pore
volume
– time step lengths might be minutes or
even seconds

Problem
• Fully Implicit is expensive
– unconditionally stable
– no limit to time step - except for time
Schlumberger Public
truncation error
– must solve for all unknowns P, Sw, Sg, in
all grid blocks simultaneously with an
iterative routine since problem is very
non-linear

Problem
• In E100 - Black Oil Simulator have a
choice of solution techniques
– IMPES - implicit pressure - explicit
Schlumberger Public
saturations
– Fully Implicit
– Lets compare IMPES and Fully Implicit

methods

IMPES – FULLY IMPLICIT
• Equations written in Residual form:

IMPES
M t + dt − M t
R= + F ( Pt + dt , St ) + Q( Pt + dt , St )
Schlumberger Public
dt
where
dM is the mass, per unit surface density, accumulated
during the current time step, dt
F is the net flow rate into neighboring grid blocks
Q is the net flowrate into wells during the timestep
R is the residual in each cell and each fluid

• IMPES residual – similar to fully

implicit residual except:
– All flow and well terms – computed
Schlumberger Public
using saturations (or Rs, Rv) AT THE
BEGINNING OF EACH TIMESTEP
• IMPES: mass terms Mt+dt – evaluated

using both pressure and saturations
at the end of the timestep

• To solve IMPES equations correctly:

– Necessary to iterate until all residuals
have been reduced to a small value
Schlumberger Public
• IMPES – linear equations from
Newton’s method – much easier to
solve – because derivative of the
flows with respect to saturations are
zero.

• IMPES: linear equations are solved

sequentially
– First for pressure
Schlumberger Public
– Subsequently for saturation
• FULLY IMPLICIT – linear equation are

solved simultaneously

IMPES and Implicit in Black Oil
 P1   P1 
   
 P2 
   Sw1 
 P3   
   Sg1 
 P4   
   P2 
 P5   
   Sw2 
Schlumberger Public
 P6   
Vector of   Vector of  Sg 2 
 P7   
Unknowns for   Unknowns for  P3 
 P8  Full Implicit  
IMPES Solution    Sw3 
 P9  Solution  
Black Oil and    Sg 3 
 P10   
   P4 
Compositional  •   
   • 
 •   
 •   • 
P   • 
 nb − 2  P 
 Pnb −1   nb 
   Swnb 
 Pnb   
 Sg nb 

Black Oil Simulation
Computational Cost Partitioning
Pie 1
The Rest
Schlumberger Public
Inversion

Dimensionality of the Linear Problem
• Given grid with 50,000 blocks:

Dimensionality
Schlumberger Public
• IMPES (Black Oil or Compositional)--
50,000
• Full Implicit 3 Phase Black Oil--
150,000
• Full Implicit 3 Phase 8 Component
• Compositional-- 500,000

Selection of Time Step
• Why needed?
– To control numerical dispersion from time
truncation.
– Accurate calculation of fluid mobilities -
Schlumberger Public
important especially for explicit systems
– To avoid dramatic changes in pressure and
saturation due to uncontrollably large time
steps

Time Stepping
• In numerical simulators
– Can specify maximum time step
– Can specify minimum time step
Schlumberger Public
– Can specify reporting steps
• change a well rate
• workover a well, add wells
• report results
– DO NOT specify the exact time step
(unless at max or min) - exact time step
controlled by commuter program

ECLIPSE Time Stepping
• TSTEP = actually reporting step

– TSTEP = 10*365.25 reporting every year
Schlumberger Public
• DATES
– DATES = 1 JAN 2001 - reporting on that
date

Internal Time Step Controls –
depending on solution approach
• Changes in pressure
• Changes in saturation
Schlumberger Public
• Material Balance errors
• Time truncation error
• Throughput ratio

Control by Pressure
• For IMPES models, 30 to 50 psi per time
step restriction is a good practice.
•
Schlumberger Public
• For fully implicit, this may be relaxed due
to unconditional stability that it offers.
• However, one must be careful to avoid

excessive numerical dispersion (time
truncation error) due to the use of large
time steps.

Control by Saturation Change
• For explicit (IMPES) schemes, the

permitted saturation change per time
step is about 5 to 10%.
Schlumberger Public
• For adaptive implicit, and fully
implicit cases, it can be further
relaxed.

Eclipse Time Step Control
Schlumberger Public
Time stepping control notes in
Eclipse 100
Schlumberger Public
Time stepping control notes in
Eclipse 100
Schlumberger Public
Special Time Stepping control
notes in ECLIPSE 100
Schlumberger Public
Reservoir Simulator Minimum
Schlumberger Public
Data Required

Minimum Data Requirements –1
• 1. Cell Values:
• Cell Dimensions ∆x, ∆y, ∆z
Schlumberger Public
• Permeability Kx, Ky, Kz
• Elevations ∆n
• Porosity φ
• Initial Pressures Pi
• Initial Saturations So, Sw, Sg
• Net to Gross ∆z net = ∆z gross x NTG

Minimum Data Requirements - 1
2. Rock and Fluid Properties
Stock Tank Densities Pos, Pws, Pgs
Fluid Viscosity µ = f(P)
Formation Volume Factors B = f(P)
Schlumberger Public
Relative Permeability Kr = f(S)
Capillary Pressure Pc = f(S)
Rock Compressibility Cf
Well Production/Controls:
Observed Rates qos, qws, qgs

Schlumberger Public
Well Calculations

What do we have to know? - An
Overview
• Well types
– Gas Injection Wells
– Water Injection Wells
Schlumberger Public
– Oil Injection Wells
– Multi-phase Production Wells
– Single Phase Oil Production Wells
– Single Phase Gas Production Wells
– Single Phase Water Production Wells

What do we have to know? - An
Overview
• The .possible modes of control
– Constant bottom hole pressure
– Constant tubing head pressure
Schlumberger Public
– Constant oil flow rate
– Constant water flow rate
– Constant liquid flow rate
– Constant gas flow rate
– Constant reservoir volume control rate

Simulator well representations
Schlumberger Public
Well Model
Field - radial flow kroKh(pe - pwf )
qo =
µo Bo [ ln( re / rw )+ Sfield ]
Schlumberger Public
simulator
kroi Ki ∆ Zi NTGi ( poi - pwf )
qo =
( µo Bo )i [ ln ( ro / rw )+ Smodel ]
grid block i
ro = Peaceman Radius

Well Model
• Well Flow Equation
 k ro 
Schlumberger Public
qoi = (WBC)i   ( pi - pwf i )

 µo Bo i

Well Model Controls
During History: qoi known, solve for Pwf
During Predictions: Pwf known, solve for qoi
Identify Well Type
Schlumberger Public
Oil Rate - qos
Gas Rate - qgs
Water Rate - qws
Liquid Rate - qos + qws
Reservoir Volume - qos Bo + qws Bw + qgs Bg

Pressure Equivalent Radius
• Following Peaceman's results, the

following equations are used in
current reservoir simulators:
Schlumberger Public
1
 1 1
 2
 k y  2  kx  2
2

  ∆x +   ∆y 
2 
 k x   ky  
ro = 0.28  1 1

 k y  4  kx  4
  +  
k k 
 x  y

Pressure Equivalent Radius
• When kx = ky, then
( )
Schlumberger Public
r o = 0 . 14 ∆ x 2
+ ∆y 2
• and if kx = ky and ∆x = ∆y,
• Then ro = 0.198 ∆x

Extended Demonstration of
Problems Involved in Subdividing
Schlumberger Public
Reservoir Into “Large” Grid
Blocks

Approximations
• In a grid block – have one value of

• Porosity
• Permeability
Schlumberger Public
• Pressure
• Oil, water, gas saturations
• Assume that have Thermodynamic

equilibrium of phases inside the grid
block

Approximations
• What does it mean to simulate a
heterogeneous piece of rock with a single
grid block?
Schlumberger Public
Following slides show the effect.
This is the situation we have in reservoir

simulation.
We must understand this situation and

understand scale-up (tomorrows lecture).

Heterogeneity for our test model
• Assume relationship between φ and k

as in sandstone of Brent Field (North
Sea) – see next slide
Schlumberger Public
• Assume porosity range is
0.15 ≤ φ ≤ 0.3
• Stochastic porosity has uniform

density in this range

Schlumberger Public
Excel Plot
100000
10000
Permeability
1000
Schlumberger Public
100
10
1
0.1 0.15 0.2 0.25 0.3 0.35
Porosity
( 26.6667φ − 4 )
k = 10
Example of Array of Porosity Values –
Random – Uniform Density
0.150 0.181 0.288 0.222 0.174 0.162 0.274 0.217 0.201 0.271
0.173 0.280 0.163 0.156 0.254 0.249 0.178 0.176 0.214 0.263
0.260 0.271 0.212 0.161 0.186 0.160 0.267 0.163 0.169 0.258
0.244 0.283 0.236 0.191 0.264 0.211 0.165 0.284 0.183 0.252
0.262 0.197 0.214 0.274 0.212 0.257 0.286 0.250 0.153 0.248
0.227 0.236 0.166 0.291 0.292 0.237 0.257 0.253 0.253 0.180
0.282 0.210 0.293 0.175 0.262 0.185 0.291 0.208 0.251 0.205
Schlumberger Public
0.281 0.201 0.179 0.253 0.170 0.207 0.199 0.228 0.159 0.285
0.158 0.275 0.241 0.160 0.210 0.229 0.222 0.290 0.163 0.186
0.295 0.251 0.160 0.196 0.157 0.272 0.278 0.195 0.299 0.161
0.180 0.275 0.162 0.274 0.266 0.284 0.243 0.257 0.228 0.244
0.282 0.213 0.208 0.247 0.169 0.181 0.289 0.279 0.216 0.167
0.258 0.283 0.207 0.227 0.283 0.160 0.193 0.292 0.265 0.171
0.184 0.264 0.152 0.226 0.211 0.250 0.167 0.196 0.154 0.250
0.258 0.167 0.207 0.215 0.249 0.218 0.194 0.244 0.158 0.165
0.230 0.184 0.224 0.181 0.217 0.209 0.238 0.182 0.197 0.279
0.282 0.269 0.262 0.212 0.223 0.229 0.256 0.234 0.205 0.257
0.190 0.261 0.252 0.294 0.279 0.193 0.155 0.151 0.161 0.241
0.175 0.203 0.264 0.234 0.276 0.176 0.222 0.246 0.215 0.177
0.161 0.297 0.169 0.258 0.214 0.248 0.154 0.291 0.166 0.206
0.184 0.192 0.157 0.163 0.177 0.160 0.284 0.227 0.181 0.219
0.188 0.165 0.252 0.239 0.207 0.230 0.256 0.177 0.234 0.185

Example of Array of Permeability
Values – Correlated to φ
1.0 6.5 4750.7 83.6 4.3 2.1 1999.4 62.7 23.4 1713.9
4.0 2901.6 2.3 1.4 576.9 438.8 5.4 5.1 50.9 1052.1
834.8 1704.8 43.7 2.0 8.9 1.8 1356.5 2.2 3.2 742.5
313.8 3573.5 202.4 12.8 1121.5 41.9 2.4 3637.4 7.5 521.8
994.9 17.4 51.9 1994.9 45.6 712.1 4329.9 456.2 1.2 411.2
Schlumberger Public
113.5 199.8 2.7 5610.8 6171.8 203.4 731.9 569.1 549.8 6.5
3332.0 39.3 6561.8 4.7 990.2 8.3 5852.4 35.3 493.0 29.4
3147.2 23.5 6.0 541.7 3.4 32.5 20.6 117.6 1.8 4073.6
1.6 2097.8 266.8 1.9 39.0 131.0 84.4 5510.8 2.3 9.1
7543.4 503.2 1.8 16.9 1.6 1794.3 2658.7 16.3 9264.2 1.9
6.2 2094.3 2.1 1978.1 1246.5 3720.1 306.0 697.6 123.1 313.1
3315.1 49.1 35.0 391.8 3.3 6.7 5155.6 2700.2 58.1 2.8
744.8 3550.5 33.6 116.4 3539.5 1.9 13.6 6225.8 1165.5 3.6
8.1 1108.0 1.1 106.2 43.0 456.8 2.8 17.3 1.3 458.6
764.1 2.8 33.4 53.0 430.4 66.0 14.5 315.3 1.7 2.6
139.4 8.2 94.2 6.6 62.0 38.5 223.8 7.2 17.7 2697.9
3311.4 1481.3 972.2 46.3 87.0 130.7 671.8 178.2 29.7 699.8
11.9 910.1 516.4 7023.6 2740.7 14.3 1.4 1.0 2.0 271.5
4.6 25.1 1079.7 172.6 2267.7 5.0 80.8 368.1 55.6 5.3
1.9 8384.4 3.2 740.2 51.2 405.7 1.2 5598.2 2.7 30.5
8.3 13.4 1.5 2.3 5.3 1.9 3745.3 113.4 6.7 67.7
10.3 2.5 521.8 234.5 33.2 139.2 683.8 5.4 168.9 8.7
Flow Field – Rock Subdivided into
Fine Grid
• Areal grid (nz = 1) – in the real world we
would have a full 3-D piece of rock
Schlumberger Public
• x and y dimensions are the horizontal
dimensions of the coarse grid block –
actually dimensions can be varied since
results will scale
• Rock contains 10,000 random

heterogeneities – size x/100 and y/100

Color Scales of Porosity and
Permeability
Schlumberger Public
Porosity
Permeability

Porosity
Schlumberger Public
Flow into block Flow out of block
Schlumberger Public
Permeability Rock boundary / coarse grid block boundary
Flow Situation
• 10,000 are represented by one grid block
in the “fine grid”
• Oil viscosity = 1 cp
Schlumberger Public
• Injection fluid viscosity (3 cases) = 1, 0.2
and .05 cp
• Relative permeabilities are straight lines –

corner to corner
• Flow into LHS of block – flow out of RHS
• Saturations shown at 5 times

Schlumberger Public
Heterogeneous Description
10,000 Blocks
Mobility Ratio = 1.0

Step 1
Schlumberger Public
Step
Step 2
Schlumberger Public
Step 2
Step 3
Schlumberger Public
Step 3
Step 4
Schlumberger Public
Step 4
Step 5
Schlumberger Public
Step 5
Schlumberger Public
10,000 Blocks

Step 1
Schlumberger Public
Step 1
Step 2
Schlumberger Public
Step 2
Step 3
Schlumberger Public
Step 3
Step 4
Schlumberger Public
Step 4
Step 5
Schlumberger Public
Step 5
Schlumberger Public
10,000 Blocks

Step 1
Schlumberger Public
Step 1
Step 2
Schlumberger Public
Step 2
Step 3
Schlumberger Public
Step 3
Step 4
Schlumberger Public
Step 4
Step 5
Schlumberger Public
Step 5
Homogeneous Case
• 10,000 heterogeneous grid blocks

are made homogeneous
Schlumberger Public
• Permeability and porosity are
average values of the 10,000 cells
• k = 1096.85 mD
• φ = 0.225

Schlumberger Public
Homogeneous Description
10,000 Blocks

Step 1
Schlumberger Public
Step 1
Step 2
Schlumberger Public
Step 2
Step 3
Schlumberger Public
Step 3
Step 4
Schlumberger Public
Step 4
Step 5
Schlumberger Public
Step 5
Single Grid Block
• For the full field simulation this
heterogeneous piece of rock is
represented by one grid block with
Schlumberger Public
– k = 1096.85 mD
– φ = 0.225
• For our simulation, small grid blocks

are connected to the single block on
the LHS and RHS for to connect an
injector and producer.
Schlumberger Public
Homogeneous Description
1 Grid Block

Step 1
Schlumberger Public
Step 1
Step 2
Schlumberger Public
Step 2
Step 3
Schlumberger Public
Step 3
Step 4
Schlumberger Public
Step 4
Step 5
Schlumberger Public
Step 5
Comparison of BT and Production
Schlumberger Public
of Injected Fluid for the Cases

Fraction of “Injection” Fluid Produced at
Right Hand Edge – Various Mobility Ratios
M = 20
Schlumberger Public
M=5
M=1
Dimensionless Time
Fraction of “Injection” Fluid Produced at Right Hand
Edge – Unit Mobility Ratios
Schlumberger Public
Single grid
block
10,000
heterogeneous 10,000
grid block homogeneous
grid block
Dimensionless Time
Conclusions
• The replacement of a very

heterogeneous section of rock with a
single “homogeneous” grid block
Schlumberger Public
where equilibrium is assumed in that
block – gives poor results in at
displacement
• BUT this is what we do in reservoir
simulation.

Schlumberger Public
End of Lecture 1

ARS-Day1 1of 3

Uploaded by

Copyright:

Available Formats

ARS-Day1 1of 3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ARS-Day1 1of 3

Uploaded by

Copyright:

Available Formats

What are the objectives of the course?

What are the objectives of the course?

What topics will be covered on day 1?

What topics will be covered on day 1?

Applied Reservoir Simulation

Reservoir Simulation Application Training

March 06 Applied Reseervoir Simulation Day 1 1

March 06 Applied Reseervoir Simulation Day 1 2

March 06 Applied Reseervoir Simulation Day 1 3

• Demonstrate key factors in reservoir

March 06 Applied Reseervoir Simulation Day 1 4

March 06 Applied Reseervoir Simulation Day 1 5

March 06 Applied Reseervoir Simulation Day 1 6

March 06 Applied Reseervoir Simulation Day 1 7

March 06 Applied Reseervoir Simulation Day 1 8

• 3. Problem 3: History Matching

March 06 Applied Reseervoir Simulation Day 1 9

• New Names for the ECLIPSE

March 06 Applied Reseervoir Simulation Day 1 11

March 06 Applied Reseervoir Simulation Day 1 12

Finite Element Mesh ,t)

Finite Difference Mesh

March 06 Applied Reseervoir Simulation Day 1 20

March 06 Applied Reseervoir Simulation Day 1 21

• What data does it need

March 06 Applied Reseervoir Simulation Day 1 22

March 06 Applied Reseervoir Simulation Day 1 23

March 06 Applied Reseervoir Simulation Day 1 24

March 06 Applied Reseervoir Simulation Day 1 26

March 06 Applied Reseervoir Simulation Day 1 27

March 06 Applied Reseervoir Simulation Day 1 28

RELATES PRESSURE/PRODUCTION with the

Solution Gas Expansion

Rock & Connate Water Expansion

Net Water Influx Production Injection

March 06 Applied Reseervoir Simulation Day 1 29

Solution Gas/Secondary Gas Cap:

Compaction Drive: Unconsolidated sands,

Water Drive: Aquifer influx due to

G ENE RAL M AT ERIAL BA LANCE E Q UATIO N

Oil Ap pro xim at es res ervo ir

March 06 Applied Reseervoir Simulation Day 1 31

Fractional Flow Concept

Welqe Tangent Method

March 06 Applied Reseervoir Simulation Day 1 34

March 06 Applied Reseervoir Simulation Day 1 35

March 06 Applied Reseervoir Simulation Day 1 37

March 06 Applied Reseervoir Simulation Day 1 38

March 06 Applied Reseervoir Simulation Day 1 43

• The following will show the workflow and

March 06 Applied Reseervoir Simulation Day 1 44

• What do we have at the start of this process?

March 06 Applied Reseervoir Simulation Day 1 45

March 06 Applied Reseervoir Simulation Day 1 46

• What do we have at the start of this

March 06 Applied Reseervoir Simulation Day 1 47

March 06 Applied Reseervoir Simulation Day 1 48

• FloGrid: From the Static Model-

March 06 Applied Reseervoir Simulation Day 1 49