PVTi REFERENCE MANUAL PDF

PVTi
Reference Manual
2005A
Proprietary Notice
Copyright © 1996-2005 Schlumberger. All rights reserved.
No part of the "PVTi Reference Manual" may be reproduced, stored in a retrieval system, or translated in any form or by any means,
electronic or mechanical, including photocopying and recording, without the prior written permission of Schlumberger.
Use of this product is governed by the License Agreement. Use of this product is governed by the License Agreement. Schlumberger.
Schlumberger makes no warranties, express, implied, or statutory, with respect to the product described herein and disclaims without
limitation any warranties of merchantability or fitness for a particular purpose.
Patent information
Schlumberger ECLIPSE reservoir simulation software is protected by US Patents 6,018,497, 6,078,869 and 6,106,561, and UK Patents
GB 2,326,747 B and GB 2,336,008 B. Patents pending.
Service mark information

The following are all service marks of Schlumberger:
The Calculator, Charisma, ConPac, ECLIPSE 100, ECLIPSE 200, ECLIPSE 300, ECLIPSE 500, ECLIPSE Office, EDIT, Extract, Fill,
Finder, FloGeo, FloGrid, FloViz, FrontSim, GeoFrame, GRAF, GRID, GridSim, Nodal, NWM, Open-ECLIPSE, PetraGrid, PIPESIM,
PIPESIM FPT, PIPESIM GOAL, PlanOpt, Prodman, Pseudo, PVTi, RTView, SCAL, Schedule, SimOpt, VFPi, Weltest 200.
Trademark information
Silicon Graphics and IRIX are registered trademarks of Silicon Graphics, Inc. OpenGL® and the oval logo are trademarks or registered
trademarks of Silicon Graphics, Inc. in the United States and/or other countries worldwide. OpenInventor and WebSpace are trademarks of
Silicon Graphics, Inc. IBM, AIX and LoadLeveler are registered trademarks of International Business Machines Corporation. Sun, SPARC,
Solaris, Ultra and UltraSPARC are trademarks or registered trademarks of Sun Microsystems, Inc. Macintosh is a registered trademark of
Apple Computer, Inc. UNIX is a registered trademark of UNIX System Laboratories. Motif is a registered trademark of the Open Software
Foundation, Inc. The X Window System and X11 are registered trademarks of the Massachusetts Institute of Technology. PostScript and
Encapsulated PostScript are registered trademarks of Adobe Systems, Inc. OpenWorks and VIP are registered trademarks of Landmark
Graphics Corporation. Lotus, 1-2-3 and Symphony are registered trademarks of Lotus Development Corporation. Microsoft, Windows,
Windows NT, Windows 95, Windows 98, Windows 2000, Windows XP, Internet Explorer, Intellimouse and PowerPoint are either registered
trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. Netscape is a registered trademark of
Netscape Communications Corporation. AVS is a registered trademark of AVS Inc. ZEH is a registered trademark of ZEH Graphics
Systems. Ghostscript and GSview are Copyright of Aladdin Enterprises, CA. GNU Ghostscript is Copyright of the Free Software Foundation,
Inc. Linux is Copyright of the Free Software Foundation, Inc. IRAP is Copyright of Roxar Technologies. LSF is a registered trademark of
Platform Computing Corporation, Canada. VISAGE is a registered trademark of VIPS Ltd. Cosmo is a trademark and PLATINUM technology
is a registered trademark of PLATINUM technology, inc. PEBI is a trademark of Veritas DGC Inc./HOT Engineering GmbH. Stratamodel is
a trademark of Landmark Graphics Corporation. GLOBEtrotter, FLEXlm and SAMreport are registered trademarks of GLOBEtrotter
Software, Inc. CrystalEyes is a trademark of StereoGraphics Corporation. Tektronix is a registered trade mark of Tektronix, Inc. GOCAD
and JACTA are trademarks of T-Surf. Myrinet is a trade name of Myricom, Inc. This product may include software developed by the Apache
Software Foundation (http://www.apache.org). Copyright (c) 1999-2001 The Apache Software Foundation. All rights reserved. MPI/Pro is a
registered trademark of MPI Software Technology, Inc. The TGS logo is a trademark of TGS, Inc. POSC, the POSC logo and Epicentre are
registered trademarks of Petrotechnical Open Standards Consortium, Inc. Red Hat is a registered trademeak of Red Hat, Inc. This product
may include software developed using LAPACK (http://www.netlib.org/lapack/), which is copyright of its authors.
Table of Contents - PVTi Reference Manual 2005A
Table of Contents - PVTi Reference Manual 2005A...............................................................................................................3
List of Figures ..... ...................................................................................................................................................................7
List of Tables ...... ...................................................................................................................................................................8
Chapter 1 - New developments...................................................................................................... 11

Developments for 2005A ......................................................................................................................................................11
Developments for 2004A ......................................................................................................................................................12
Chapter 2 - The Most Asked Questions About PVTi.................................................................... 15

Introduction ......... .................................................................................................................................................................15
Chapter 3 - Introduction ................................................................................................................. 27

General information ..............................................................................................................................................................27
Chapter 4 - Getting started ............................................................................................................ 31

Starting PVTi ...... .................................................................................................................................................................31
Chapter 5 - Tutorials ....................................................................................................................... 33

Overview............. .................................................................................................................................................................33
Fluid Properties Estimation...................................................................................................................................................35
Creating a fluid system .........................................................................................................................................................39
Simulating experiments ........................................................................................................................................................45
Fitting an equation of state to experimental results ..............................................................................................................53
Exporting ECLIPSE Black Oil PVT tables.............................................................................................................................59
Converting a black oil run to compositional ..........................................................................................................................65
Workflow Tutorial .................................................................................................................................................................69
Multiphase Flash .................................................................................................................................................................77
Exporting an ECLIPSE Thermal model ................................................................................................................................81
Data analysis and quality control ..........................................................................................................................................85
Removing contamination from samples................................................................................................................................93
Converting old projects to the current version ......................................................................................................................97
Chapter 6 - Reference section ....................................................................................................... 99

General information ..............................................................................................................................................................99
Main PVTi window ..............................................................................................................................................................100
The PVTi main module .......................................................................................................................................................101
The fluid model ... ...............................................................................................................................................................109
COMB - Compositional Material Balance ...........................................................................................................................123
Simulation using PVTi ........................................................................................................................................................128
Regression in PVTi .............................................................................................................................................................137
Exporting keywords ............................................................................................................................................................144
VFP module ........ ...............................................................................................................................................................149
Utilities ................ ...............................................................................................................................................................155
Batch system and keywords ...............................................................................................................................................163
Error handling ..... ...............................................................................................................................................................174
Chapter 7 - Keywords ................................................................................................................... 177

PVTi keywords.... ...............................................................................................................................................................177
Keywords A-D..... ...............................................................................................................................................................178
ACF Acentric factors ..............................................................................................................................................179
ACHEUH A-coefficient for Cheuh-Prausnitz BICs .........................................................................................................180
3
ALLDRY Dry Gas Tables for Each Sample ...................................................................................................................181
BIC Binary interaction coefficients ........................................................................................................................182
BLACKOIL Start of the BLACKOIL section ......................................................................................................................184
CALVAL Specify calorific values ..................................................................................................................................185
CHARACT Components to be characterized...................................................................................................................186
CNAMES Component names ........................................................................................................................................187
COATS Blackoil tables ................................................................................................................................................188
COMB Start of the COMB section .............................................................................................................................189
COMBINE Group existing components...........................................................................................................................190
CORRACF Splitting correlation for ACFs.........................................................................................................................191
CORRCP Splitting correlation for critical properties.......................................................................................................192
DRYGAS Dry gas tables..................................................................................................................................................193
DEADOIL Dead oil tables...............................................................................................................................................194
DEBUE Select output to debug file .............................................................................................................................195
DEBUG Select output to debug file .............................................................................................................................196
DEFBIC Default binary interaction coefficients ............................................................................................................197
DEGREES Temperature convention................................................................................................................................198
DIFFERENTIAL Blackoil tables........................................................................................................................................199
DREF Reference densities .......................................................................................................................................200
Keywords E-K ..... ...............................................................................................................................................................201
ECHO Insert PVI file into PVP file ............................................................................................................................202
EOS Defines the required Equation of State ..........................................................................................................203
EOSOUT EoS data for ECLIPSE 300...........................................................................................................................204
EXP Experiments ..................................................................................................................................................205
EXPIND Set Status of Experiments ................................................................................................................................210
FIT Perform fit by regression ...............................................................................................................................211
FRAC Specify plus fraction data ..............................................................................................................................212
FRAGOR Blackoil tables................................................................................................................................................213
FVFREF FVF reference conditions...............................................................................................................................214
GI Define GI nodes for E200 tables .....................................................................................................................215
GROUP Start of the GROUP section...........................................................................................................................216
GRBYALL Start of the GROUP section ..........................................................................................................................217
GRBYMIX Start of the GROUP section ..........................................................................................................................218
GRBYSAM Start of the GROUP section ..........................................................................................................................219
GRPBYWGT Grouping by molecular weight .......................................................................................................................220
HYDRO Define component as hydrocarbon or non-hydrocarbon................................................................................221
KVTABLE Request K-value table for ECLIPSE 300 output............................................................................................222
Keywords L- O .... ...............................................................................................................................................................223
LBC Lohrenz-Bray-Clark viscosities.......................................................................................................................224
LBCCOEF Set non-default LBC coefficients ...................................................................................................................225
LIVEOIL Live oil tables.................................................................................................................................................226
LNAMES Specify library names ....................................................................................................................................227
MAXIT Max. number of regression iterations.............................................................................................................228
MAXSTEP Maximum step size allowed in regression .....................................................................................................229
MDP Data for Whitson splitting ...............................................................................................................................230
MESSAGE Echo message to file and screen ..................................................................................................................231
MINDELP Minimum pressure difference ........................................................................................................................232
MINSTEP Minimum step limit allowed in regression ......................................................................................................233
MIX Mix samples ...................................................................................................................................................234
MODSPEC Denotes start of the run specification section ................................................................................................235
MODSYS Start of the MODSYS section ........................................................................................................................236
MOSES Blackoil tables ................................................................................................................................................237
MW Specify molecular weights..............................................................................................................................238
MWS Define plus fraction mole weight for CMF splitting .........................................................................................239
NCOMPS Specify number of components .....................................................................................................................240
NEWPVI Request new output PVI file ..........................................................................................................................241
NEWPVO Request new output PVO file.........................................................................................................................242
NOECHO No insertion of PVI file into PVP file...............................................................................................................243
OBS Specify observations ......................................................................................................................................244
4
OBSIND Specify observation weights ..........................................................................................................................245
OMEGAA,OMEGAB
Specify EoS omega values.................................................................................................................................................247
OPTIONS Set various program options ........................................................................................................................248
OUTECL3 Start of the OUTECL3 section.......................................................................................................................250
Keywords P- S .... ...............................................................................................................................................................251
PARACHOR Define parachors ..........................................................................................................................................252
PCRIT Critical pressures ...........................................................................................................................................253
PEARCE Blackoil tables................................................................................................................................................254
PEDERSEN Specify Pedersen viscosities........................................................................................................................255
PRCORR Peng-Robinson correction .............................................................................................................................256
PSEUCOMP Start of the PSEUCOMP section...................................................................................................................257
RECOVERY Liquid production for recovery estimates ......................................................................................................258
REGRESS Start of the REGRESS section......................................................................................................................259
REGTARG Regression target ..........................................................................................................................................260
RTEMP Reservoir temperature for ECLIPSE Compositional......................................................................................261
RUNSPEC Denotes start of the run specification ............................................................................................................262
SALINITY Specify sample salinity .................................................................................................................................263
SAMPLE Specify fluid sample.........................................................................................................................................264
SAMPLES Specify fluid samples.....................................................................................................................................265
SAMPLES Specify fluid samples.....................................................................................................................................266
SAMTITLE Specify titles of fluid samples ........................................................................................................................267
SAVCOMP Save compositions ........................................................................................................................................268
SCT Defines Semi-Continuous Thermodynamics split ...........................................................................................269
SG Specify specific gravity...................................................................................................................................270
SIMULATE Start of the SIMULATE section .....................................................................................................................271
SPECHA,B,C,D
Specify specific heat capacity coefficients..........................................................................................................................272
SPLIT Start of the SPLIT section..............................................................................................................................273
SSHIFT Dimensionless volume shifts for PR3 ............................................................................................................274
STCOND Standard conditions.......................................................................................................................................275
SYSTEM Start of the SYSTEM section.........................................................................................................................276
Keywords T - Z ... ...............................................................................................................................................................277
TBOIL Specify boiling points .....................................................................................................................................278
TCRIT Specify critical temperatures...........................................................................................................................279
THERMX Thermal expansion coefficient for volume shifts............................................................................................280
TITLE Specify run title ..............................................................................................................................................281
TLOW Define lowest temperature for VFP tables ......................................................................................................282
TREF Specify reference temperatures.....................................................................................................................283
UNITS Specify unit conventions ................................................................................................................................284
VAR Specify regression variables ..........................................................................................................................285
VCRIT Specify volumes.............................................................................................................................................289
VCRITVIS Specify volumes for LBC viscosity calculations ...........................................................................................290
VERSION Version of PVTi .............................................................................................................................................291
VFP Start of the VFP section .................................................................................................................................292
WAT100 Output water properties .................................................................................................................................293
WATVFP Output water properties .................................................................................................................................296
WETGAS Wet gas tables.................................................................................................................................................297
WHIT Defines Whitson splitting.................................................................................................................................298
WHITSON Blackoil tables ...............................................................................................................................................299
XMFVP/YMFVP
............................ XMFVP and YMFVP ECLIPSE tables ...................................................................................................300
ZCRIT Specify critical Z-factors.................................................................................................................................301
ZCRITVIS Specify critical Z-factors for LBC calculations ...............................................................................................302
ZI Specify sample composition...........................................................................................................................303
ZMFVD Composition versus depth table ....................................................................................................................304
5
Chapter 8 - Technical Description ............................................................................................... 305
Overview ............. ...............................................................................................................................................................305
Theoretical background of PVT ..........................................................................................................................................306
Equation of state . ...............................................................................................................................................................330
Basic laboratory experiments..............................................................................................................................................355
Regression ......... ...............................................................................................................................................................365
Output for ECLIPSE simulators ..........................................................................................................................................372
Analysis techniques ............................................................................................................................................................392
Recommended PVT analysis for oil reservoirs ...................................................................................................................393
Recommended PVT analysis for gas condensate reservoirs .............................................................................................398
Consistency tests and correlations .....................................................................................................................................403
Fluid Properties Estimation .................................................................................................................................................406
Regression in PVT analysis ................................................................................................................................................408
Wax and asphaltene precipitation in PVTi ..........................................................................................................................416
Cleaning samples contaminated with oil-based mud..........................................................................................................420
Mixing and recombination of samples.................................................................................................................................422
ECLIPSE Thermal Export Module ......................................................................................................................................423
Appendix A - Units........................................................................................................................ 431

Units.................... ...............................................................................................................................................................431
Appendix B - Symbols.................................................................................................................. 435

Symbols .............. ...............................................................................................................................................................435
Bibliography
Index
6
List of Figures
Chapter 2 - The Most Asked Questions About PVTi.................................................................... 15

Figure 5.1 Fingerprint Plot .......................................................................................................................................42
Figure 5.2 Phase Plot ..............................................................................................................................................43
Figure 5.3 The plotted simulation results .................................................................................................................48
Figure 5.4 Plot of Oil FVF, Viscosity and Rs versus pressure for the output black oil property tables ....................61
Figure 5.5 Phase Diagram for Schrader Bluff Fluids ...............................................................................................78
Figure 5.6 The phase envelope plot. .......................................................................................................................86
Figure 5.7 The main display shows messages indicating the quality of the data.....................................................87
Figure 5.8 The main plot window after zooming in ..................................................................................................88
Figure 5.9 The plot of k values versus pressure. .....................................................................................................89
Figure 5.10 The Hoffman-Crump plot ........................................................................................................................90
Figure 5.11 Hoffman-Crump-Hocott plot. ...................................................................................................................91
Figure 5.12 The original sample, the cleaned sample and the estimated contaminant. ............................................94

Figure 6.1 The main PVTi window .........................................................................................................................101
Figure 6.2 Fingerprint Plot .....................................................................................................................................120
Figure 6.3 Phase plot .............................................................................................................................................121
Figure 6.4 Ternary Plot .........................................................................................................................................122
Figure 6.5 Main display after performing material balance ....................................................................................124
Figure 6.6 COMB module - vapor versus pressure plot ........................................................................................125
Figure 6.7 The VFP module ...................................................................................................................................149
Appendix A - Units........................................................................................................................ 431
Bibliography
7
List of Tables
Chapter 2 - The Most Asked Questions About PVTi ................................................................... 15

Table 5.1 The Fundamentals panel .........................................................................................................................36
Table 5.2 Component and fluid definitions...............................................................................................................40
Table 5.3 Program Options data ..............................................................................................................................41
Table 5.4 Constant Composition Expansion experiment at 220o F (* indicates bubble point pressure)..................46
Table 5.5 Differential Liberation Experiment at 220o F (* indicates bubble point pressure).....................................49

Table 6.1 List of library components ......................................................................................................................105
Table 6.1 Observation data....................................................................................................................................134
Table 6.2 Set PVTi Program Options panel ...........................................................................................................156
Table 6.3 Keywords for introducing sections .........................................................................................................167
Table 6.4 RUNSPEC keywords .............................................................................................................................167
Table 6.5 SYSTEM keywords ................................................................................................................................168
Table 6.6 SPLIT keywords .....................................................................................................................................169
Table 6.7 GROUP keywords..................................................................................................................................169
Table 6.8 COMB keywords ....................................................................................................................................170
Table 6.9 SIMULATE keywords .............................................................................................................................170
Table 6.10 REGRESS keywords .............................................................................................................................170
Table 6.11 BLACKOIL keywords .............................................................................................................................171
Table 6.12 PSEUCOMP keywords ..........................................................................................................................171
Table 6.13 OUTECL3 keywords ..............................................................................................................................172
Table 6.14 VFP keywords ........................................................................................................................................172
Table 6.15 APITRACK keywords .............................................................................................................................173
Table 6.16 Error codes ............................................................................................................................................174

Table 7.1 Output indices ........................................................................................................................................195
Table 7.2 Output indices ........................................................................................................................................196
Table 7.3 Required data for experiments...............................................................................................................205
Table 7.4 Keyword arguments ...............................................................................................................................206
Table 7.5 Restrictions for EXP keyword arguments...............................................................................................208
Table 7.6 Component Types..................................................................................................................................221
Table 7.7 Equation of State omega values ............................................................................................................247
Table 7.8 Default limits for variables ......................................................................................................................286

Table 8.1 Alkanes ..................................................................................................................................................308
Table 8.2 Napthenes..............................................................................................................................................309
Table 8.3 Aromatics ...............................................................................................................................................309
Table 8.4 Physical properties.................................................................................................................................310
Table 8.5 Multi-component (ii) mixtures .................................................................................................................311
Table 8.6 Multi-component (ii) mixtures .................................................................................................................311
8
Table 8.7 CVD Report ...........................................................................................................................................320
Table 8.8 Equation of State coefficients ................................................................................................................332
Table 8.9 Equation of State constants...................................................................................................................334
Table 8.10 Parameter estimation data. N is the number of experimental points .....................................................351
Table 8.11 Parameter Values for Pure Component Viscosity Correlation ...............................................................352
Table 8.12 Physical Properties of Methane and Decane.........................................................................................353
Table 8.13 PVTi defaults for Fluid Property Estimation ...........................................................................................407
Appendix A - Units........................................................................................................................ 431

Table A.1 Units.......................................................................................................................................................432
Table A.2 Constants...............................................................................................................................................433
Table A.3 Conversion factors .................................................................................................................................433
9
10
New developments
Chapter 1
Developments for 2005A

Maintenance of this application is continuing until further notice.
New developments 11
ECLIPSE Thermal Export facility

For the 2003A version of PVTi a new ECLIPSE Thermal support module was available
where you were able to interactively develop a correlation which accurately predicted
K-values for each component in a given fluid. For the 2004A version this module has
been extended to a full export facility where you can write out files that are suitable for
use as PVT input for ECLIPSE Thermal. The motivation behind this is so that, just as
you can export files to use as PVT input for ECLIPSE BlackOil and ECLIPSE
Compositional, they will now be able to do the same for ECLIPSE Thermal.
PVTi will export a series of keywords when an export for ECLIPSE Thermal is
performed. For a workflow description and brief summary of these keywords see
"Compositional Data for ECLIPSE Thermal" on page 387. For a more technical outline
of how the exported keywords are used in ECLIPSE Thermal see "ECLIPSE Thermal
Export Module" on page 423.
Export for API Tracking option in ECLIPSE

BlackOil
The API Tracking facility enables ECLIPSE BlackOil to model the mixing of different
types of oil, having different surface densities and PVT properties.
Without the API Tracking facility, the presence of different types of oil in the reservoir
could be handled with the aid of PVT region numbers. Oil in PVT region 1 would have
its properties determined from PVT table number 1, and so on. However, this method
cannot model the mixing of oil types. Oil flowing from region 1 into region 2 would
appear to take on the properties associated with region 2.
The API Tracking facility essentially replaces the concept of PVT regions for oil. The
PVT tables used for determining the oil properties are selected at each time step
according to the average API of the oil in each grid block (or to be more precise, its
average surface density).
For a overview of the workflow involved to export PVT tables suitable for use in
ECLIPSE BlackOil with the API Tracking option turned on see "Export for API Tracking
option in ECLIPSE BlackOil" on page 145. For a more technical description of the API
Tracking model in ECLIPSE as well as an explanation of how PVTi calculates suitable
PVT tables see "Model for API Tracking option in ECLIPSE BlackOil" on page 383.
Batch Mode
For the 2004A version of PVTi the batch mode has undergone a significant revamp.
Over the last few years the user interface of PVTi has evolved rapidly and the existing
batch mode facility no longer adequately supports more recent functionality. There
have been 3 significant modifications to the PVTi batch mode:
12 New developments
1 The way a batch mode is executed has changed. The new way to launch a batch
mode run on a PC is to use the command $pvti -batch filename where
filename is the name of your PVTi project. See "General information" on page
163 for more details on running batch mode with other platforms.
2 The new batch mode now supports the majority of the functionality available in
interactive mode - namely splitting, grouping, regression, experiment simulation
and export. Export for ECLIPSE Thermal, PNA splitting and material balance
checks are not currently supported. For more details on the functionality
constraints of the batch mode see "Constraints on the workflow" on page 165.
3 It was felt that a more user friendly way of constructing the .PVI files suitable for
batch mode was needed since a batch mode file requires some extra sections than
a standard interactive mode PVI file does. A new option called Write Keywords for
Batch Mode is available on the Options panel. If this is turned on and a workflow
is performed in interactive mode and then the file saved, PVTi ensures that this file
is then suitable for use in a batch mode run. The batch mode run then reproduces
the workflow and results that were obtained in interactive mode. See"Preparing
Batch Mode Files in Interactive Mode" on page 164 for a detailed explanation of
this facility.
For an overview of all aspects of the new batch mode facility see "Batch system and
keywords" on page 163.
Panel Addition/Modification
1 There is a new LBC Viscosity Coefficients panel available under the menu option
Edit | Fluid Model | LBC Viscosity Coefficients.... This panel shows, if using the
Lorentz-Bray-Clark (LBC) viscosity model, the current values of the five
coefficients.
2 The Thermal Properties panel has been modified to include the new properties
introduced as a result of the new ECLIPSE Thermal export functionality.
Specifically, the properties Heat Cap. A and Heat Cap. B have been changed to
Heat Cap. A/G and Heat Cap. B/H. There is also a new property called Heat of
Vaporization which corresponds to the HEATVAPS keyword exported for ECLIPSE
Thermal.
Component Constraints
1 It is now possible to read in projects where fluid samples can have up to 100
components. However, no functionality involving the EoS flash can be used until
a group operation has been performed in order to reduce the number of
components in all samples to <=50.
2 Two new components are available in the PVTi library - Benzene (C6H6) and
Toluene (C7H8). The short names for use in the fundamentals or components panel
are BEN for Benzene and TOL for Toluene. See "Library" on page 105 to view the
components in the PVTi library. To view the library within PVTi itself use the View
| Library option.
New developments 13
New Keywords and sections
Keywords
There are two new keywords in PVTi 2004A.
1 EXPIND.
This is a list of integers - one for each experiment defined in the project - which specify
whether an experiment should be used (‘turned on’) or not used (‘turned off’) when
performing a regression.
2 HEATVAPS.
This keyword is used to store the Heat of Vaporization at the standard temperature for
each component.
Sections
There is a new section called APITRACK. This is used when preparing batch files if the
API Tracking export facility is to be made use of in batch mode. It is essentially the
same as the BLACKOIL section but has an extra keyword called SAMPLES, which
records the samples for which PVT tables will be exported.
Manual
1 The section "The Most Asked Questions About PVTi" on page 15 has been updated
with questions regarding the Batch Mode, ECLIPSE Thermal Export and API
Tracking functionality.
2 A number of the tutorials have been amended - in particular the tutorial entitled
“Using the ECLIPSE Thermal Support Module” has been replaced by a tutorial
called "Exporting an ECLIPSE Thermal model" on page 81.
14 New developments
The Most Asked Questions
About PVTi Chapter 2
Introduction
This section has been designed as a reference section so that you can quickly access
information about common problems encountered with PVTi without having to spend
time looking through the manual for the relevant section. The questions in this section
have been constructed using the most common support issues and also the InTouch
database. Cross-references are provided where necessary so that readers can access the
appropriate parts of the manual for more detailed information on a particular topic if
required.
The questions are:
• "What is PVTi used for? Why do we need it?" on page 16
• "Where do I start? How do I set up a project within PVTi?" on page 16
• "How do I create an experiment along with a series of observations?" on page 17
• "What are the data limitations in PVTi?" on page 18
• "What is the Fluid Properties Estimation facility in PVTi?" on page 19
• "How do I perform regression on multiple fluid samples?" on page 19
• "What regression parameters should I choose?" on page 21
• "What is the difference between normal regression, special regression and
automatic (PVTi selects) regression?" on page 20
• "How does PVTi support gas condensate simulation?" on page 22
• "Can Black Oil tables be extended above the liquid bubble point in PVTi?" on page
23
• "Can PVTi be used if you know the composition of a fluid but do not have any
observations? And vice-versa?" on page 23
• "What black oil correlations are available in PVTi?" on page 24
• "How do I generate the asphaltene phase envelope using PVTi?" on page 24
The Most Asked Questions About PVTi 15

Introduction
• "How does PVTi support ECLIPSE Thermal?" on page 25
• "How do I Use PVTi’s Batch Mode?" on page 25
• "How Can I Export PVT Tables to use the API Tracking Functionality in ECLIPSE
BlackOil?" on page 26
What is PVTi used for? Why do we need it?

PVTi is a compositional PVT equation-of-state based program used for characterizing
a set of fluid samples for use in our ECLIPSE simulators.
We need PVTi because it is vital that we have a realistic physical model of our reservoir
fluid sample(s) before we try to use them in a reservoir simulation. PVTi can be used
to simulate experiments that have been performed in the lab on a set of fluid samples
and then theoretical predictions can be made of any observations that were performed
during a lab experiment, in order that we can test the accuracy of our fluid model.
Any differences between the measured and calculated data are minimized using a
regression facility which adjusts various Equation of State parameters. This ‘tuned’
model is then exported in a form suitable for one of our ECLIPSE simulators.
What is especially important to note when using the ECLIPSE Compositional
simulator is that PVTi and ECLIPSE Compositional use the same flash algorithm. This
is vital as the flash has been used to simulate the experiments and predict values for
experimental observations and is therefore inherent within the fluid model itself which
has been exported by PVTi. If ECLIPSE Compositional used a different flash then the
fluid model exported by PVTi would no longer be valid.
Where do I start? How do I set up a project within

PVTi?
Defining a Fluid Sample
If you want to open a new project then start PVTi as instructed in "Getting started" on
page 31 for your machine-type and choose a filename. PVTi starts; recognizes that it
has a new project and immediately opens the Fundamentals panel. This panel has been
specifically designed to make setting up a new project as easy as possible. Simply fill
in the Components and ZI columns with the component names and mole fractions
respectively, which is the minimum required to have a project within PVTi.
To fill in the component names simply type the standard shorthand names for the
components in your fluid, for example, C1, N2, CO2, H2S, IC5, etc. For more
information on shorthand names and component types see "Component types" on
page 113. The mole fractions can be entered as fractions or percentages by selecting the
appropriate option on the panel. Also, weight fractions/percentages can be entered for
the components instead of mole fractions/percentages.
Your Mole/Weight data must add up to 1 if entering as fractions and up to 100 if
entering as percentages. If they do not then PVTi asks you if you want it to renormalize
your data when you close the Fundamentals panel. If you want to add some
components then select No,- otherwise select Yes.
16 The Most Asked Questions About PVTi

Introduction
Warning If you try to perform operations on a fluid with incorrectly normalized
mole/weight fractions then the operation may fail.
Once the Fundamentals panel has been completed you will see have a sample called ZI
on the tree view on the left-hand side of the main window.This is the fundamental
sample for the project and the name ZI cannot be changed.
Creating Other Fluid Samples

Other fluid samples can be created in a project by selecting Edit | Samples | Name... .
Simply type in the name of the new fluid sample you want to create. The composition
information can then be entered for this fluid by selecting the Edit | Samples |
Compositions... option.
Additional fluid samples to the project must always be subsets of the ZI sample in
terms of the component names, for example you cannot have a C8 component in an
additional fluid sample called OIL if C8 was not defined in the ZI sample. If you open
the Edit | Samples | Compositions... panel you can see why this has to be the case in
PVTi.
Note Just because a component is defined in the ZI sample it does not mean there
has to be any of it there. It only has to be defined in the ZI sample to be used in
other fluid samples. If the mole fraction of a C8 component in the ZI sample is
set to be 0.0 then the C8 component can then be use in the OIL sample and the
mole fractions set as required.
Once at least one fluid sample (the ZI sample) has been defined then any experiment
supported within PVTi can be simulated as well as operations such as phase plots,
fingerprint plots and splitting.
For more information on creating fluid samples see "Defining Samples" on page 118.
How do I create an experiment along with a series

of observations?
Creating Experiments
To create an experiment select the Edit | Experiments... option and the experiment
Entry panel opens. The existing experiments are listed and you can edit them by
selecting one of them and clicking the next button. To create a new experiment click
add in the top left of the panel and select the experiment you wish to create. Choose
the fluid sample you want to perform the experiment on and then navigate through the
panels by filling in the required information and then clicking next, which takes you to
the next panel. The information generally consists of temperature and/or pressure
information but not always, it depends on the experiment.
Once an experiment has been created an experiment button, along with an experiment
name, appears below the fluid sample which the experiment was performed on.

Introduction
Creating Observations
If there are no observations at all for a particular experiment then to create one you
need to select the Edit | Observations... option and the Observations panel appears.
On the Experiments column on the left-hand side there is a list of all the experiments
that are available within PVTi and * symbols are next to the ones that you currently
have defined within your project.
If one of these experiments is selected then in the Experiment List column a list of all
the names of the experiments of that type in your project appears for example,
BUBBLE5, DEW3, DL1.
If one of these is selected then all the possible observations available within PVTi for
that type of experiment are displayed in the Observation Type column. Again
observation types with a * next to them means that there are values already defined for
this particular experiment in your project. Simply click on one to see and edit the
values. To create a new observation select the one you want and then click on the +
button on the top left of the panel. Values and weights can then be entered for the
observation.
Note Currently defined observations for an experiment can be edited in the

Observations folder on the experiment Entry panel.
For more information on creating and editing experiments/observations see

"Simulation using PVTi" on page 128 and/or the tutorial "Simulating experiments" on
page 45.
What are the data limitations in PVTi?

Pre-2003A
Up to and including the 2002A_1 release (pre-2003A) the following data constraints
were present in PVTi:
• 50 fluid samples
• 50 components per fluid sample* (see below)
• 50 experiments per fluid sample
• 300 observations per experiment
Note *When a splitting operation was performed it was possible to have more than
50 components (up to 100 in fact) but the components had to be grouped back
so that there were less than 50 before any experiment simulation could take
place.
2003A
These pre-2003A data constraints have been present in PVTi for 4 to 5 years and, in-line
with the huge increase in computing power in the last few years, we have decided to
enhance the data constraint capability of PVTi so that the following is now available:
• 100 fluid samples

Introduction
• 100 components per fluid sample* (see below)
• 100 experiments per fluid sample
• 300 observations per experiment
Note It is now possible to read in, save, split and group with fluids containing up to
100 components. However, the limit is still 50 components for any
functionality involving the EoS flash.
What is the Fluid Properties Estimation facility in

PVTi?
The Fluid Properties Estimation (FPE) facility in PVTi is designed so that it can be used
when you have minimal data at your disposal, at the well-site for example. In this
scenario, a full lab analysis of multiple fluid samples from the reservoir has not yet
been performed. Typically, just a single sample would be available and minimal fluid
behavior known for example, saturation pressure at a particular temperature.
Specifically, the FPE facility assumes that a single fluid sample with compositional
information is available which includes a single plus fraction (for example C7+)
component of which the weight fraction is known. Typically, this weight fraction data
is fairly accurate but the mole weight, which is used to characterize the critical
properties of the plus fraction, is not. The FPE functionality allows you to perform a
quick look simulation that regresses on the mole weight of the plus fraction, and keep
the weight fraction constant, in order to fit to a saturation pressure observation at a
particular temperature.
The FPE facility is available in the top right-hand corner of the fundamentals panel
whenever a new project is created. Alternatively it can be accessed using the Edit |
Properties Estimation (FPE)... option. For more information on this facility see "Fluid
Properties Estimation" on page 406. For an example of how it works see the tutorial
"Fluid Properties Estimation" on page 35.
How do I perform regression on multiple fluid

samples?
General
The fluid samples that PVTi performs regression on is determined by the structure of
the tree view on the left-hand side. By default, PVTi performs a regression on every
experiment which has observations defined, even if there are multiple fluid samples,
each with their own experiments. The reason for this is that, within a project, all fluid
samples are considered to be relevant to each other and so the same fluid model should
be applied to all samples, even if the compositional make-up of each sample is different
Note If two of your fluid samples are not relevant to each other for example they
come from different wells/reservoirs then a separate project should be created
for each one.

Introduction
Disabling Experiments/Observations
You can prevent PVTi from including an experiment in the regression by right-clicking
on the experiment and selecting Don’t use in Regression. A cross appears on the
experiment indicating it is not currently available within the regression facility.
You can disable an observation so that it is not used within the regression by again
right-clicking and selecting Don’t use in Regression. Alternatively, by right-clicking
and selecting Set Weight and then entering zero the observation is also not included in
the regression.
Note If an experiment is disabled then, as you would expect, all the observations are
automatically disabled.
Regression Weights
In general there will be a set of values in an observation. For example, if we have a
differential liberation (DL) experiment defined then a viscosity observation would
have a value for each pressure. We have two types of weight: there are single weights
for each value of an observation and global weights that apply to every value in an
observation.
By right clicking on an experiment observation the global weight can be set. As
mentioned above, by setting this to zero none of the values in the observation would
be used. Alternatively, you may want to set a global weight for an experimental
observation particularly high, for example, matching the bubble point of a fluid is
normally very important if one wants to ensure that it is a single-phase liquid at the
temperature and pressure of the reservoir.
Or maybe you do not trust the accuracy of a particular observation value, for example
an oil formation volume factor (FVF) value in a DL experiment.You may then not want
to use a global weight as all the other observation values look ok. In this case setting a
single-value weight to a very low value helps you match all the other values in the
observation during regression as the rogue, inaccurate value no longer inhibits
convergence.
Both the single-value and global weights for an experimental observation can be set in
the Observations panel by selecting the Edit | Observations... option, highlighting the
appropriate observation and then simply typing in your chosen weights.
For a good example of how to use the regression facility, see the tutorial "Fitting an
equation of state to experimental results" on page 53.
What is the difference between normal regression,

special regression and automatic (PVTi selects)
regression?
There are 3 types of regression: normal, special and automatic. The difference between
them depends entirely on what kind of variables are being regressed on.

Introduction
Normal regression parameters are equation of state variables relating to a particular
component, for example, critical pressure, P c , critical temperature, T c , acentric factor,
ω . and the binary coefficients. The full set of normal regression variables can be
viewed using the regression panel using the Run | Regression... panel. Select normal
as the regression type and then click variables - the upper table shows the single-
valued normal regression parameters for each component and the lower panel shows
the binary coefficients table. For more information on setting normal regression see
"Setting normal variables" on page 139.
Special regression parameters are global Equation of State variables, for example, the
thermal expansion coefficient or the Cheuh-Prausnitz A factor for binary coefficients.
There may also be some splitting parameters available as special regression variables
depending on whether a multi-feed split has been performed on the plus fraction. See
"Multi-feed Split (also called semi-continuous thermodynamic (SCT) split)" on page
117 for more details on this facility. For more information on setting special regression
variables see "Setting special variables" on page 140.
The automatic regression facility or PVTi Selects as it is called on the Regression panel
attempts to choose the best normal variables to regress on for you before actually doing
the regression and reporting the answer. It should be noted that there is no substitute
for a good engineer in the sense that one should not just use this automatic facility all
the time in the belief that PVTi will do all the work. For example, the automatic facility
will not use special variables to regress on and so it is up to you to decide whether this
would be necessary or not.
However the automatic facility can still be a useful tool in obtaining a good match to
PVT data. The way the algorithm chooses the regression parameters is essentially
based on 2 criterion. No parameters are allowed to have more than a 90% correlation
on any other parameter. Secondly, no parameter is allowed to have less than 1% of the
sensitivity of the most sensitive parameter For a detailed discussion of the automatic
regression algorithm see "A consistent methodology that can be applied automatically"
on page 411.
For more general information on the regression facility see "Regression in PVTi" on
page 137.
What regression parameters should I choose?

It should be noted that there are no concrete rules for getting a good match to
observations relating to multiple fluid samples, but there are some general guidelines
of what is often a good idea, and what you should definitely not do.
Library components tend to have properties that are very well known and any of these
will not normally be good choices of regression variables. Properties of non-library
components and characterized components are much less well known and these are
often good choices. In general, the following set of variables are normally good things
to initially regress on:
• P c, T c and ω of any non-library component.
• P c, T c and ω of any component with mole weight of C7 or heavier (as these are
effectively mixtures of different molecule types and so may differ from library
values).

Introduction
• Ω A and Ω B of any component with mole weight C7 or heavier. Again because
these are mixtures.
• No binary inter-active coefficients because of the risk of over-fitting.
• No viscosity-specific parameters, again because of the risk of over-fitting.
The variables mentioned above are all normal regression variables. The following set
of special regression variables can also often prove useful to get a match between
samples:
• Do a multi-feed split to split the plus fraction into 2 or 3 pseudo-components.
The α , SCTMW and Kw parameters, which control the splitting are then very
good choices.
• The Cheuh-Prausnitz A binary parameter if using Cheuch-Prausnitz binaries.
• The mole weight of a plus fraction (if no split on the plus fraction has been
performed).
For a detailed explanation of why some of the above are good and bad choices for
regression parameters, see "Regression in PVT analysis" on page 408. There are also
more specific guidelines for choosing regression parameters depending on whether
one is dealing with an oil reservoir, see "Recommended PVT analysis for oil reservoirs"
on page 393 - or a gas-condensate reservoir, see "Recommended PVT analysis for gas
condensate reservoirs" on page 398.
Finally, for a tutorial illustrating the use of the normal and special regression facilities
in a typical workflow see the new "Workflow Tutorial" on page 69.
How does PVTi support gas condensate

simulation?
Gas condensate simulation is modeled in PVTi using the Constant Volume Depletion
(CVD) experiment. In most gas condensate reservoirs liquid does reach a high enough
saturation to become mobile as the pressure drops. Thus, gas and oil, to a good
approximation, do not move with respect to each other and so the CVD experiment
models this behavior very well. Other experiments that tend to be used for gas
condensates are the dew point and constant composition expansion experiments. For
more information see "Gas condensate systems" on page 358.
A common failing when analyzing gas condensate reservoirs is to attempt to establish
an equation of state representation without a through examination of the data on
which it is to based. Things to check in the data are the characteristics of the heavy
components (use a fingerprint plot), material balance information and other
information such as K-values and Z-factors. Once you are happy with the data that you
will try to match there is a recommended procedure, in terms of regression, which will
work for most gas condensates. For a detailed description on how to model gas
condensates see the section entitled "Recommended PVT analysis for gas condensate
reservoirs" on page 398.

Introduction
Can Black Oil tables be extended above the liquid
bubble point in PVTi?
Yes. You should make sure that the DL/CVD experiment you simulate in PVTi covers
the full range of pressure values you are likely to encounter within your reservoir. If
ECLIPSE BlackOil encounters a pressure outside the range in the black oil table you
exported from PVTi then it will have no choice but to try to extrapolate to estimate
properties such as gas-oil ratio (GOR) and formation volume factor (FVF). The
extrapolation used is linear and uses the appropriate quantities at the two highest
pressures in the exported black oil table.
However, this extrapolation can sometimes run into difficulties as is common in any
problem when you are trying to gain information about unexplored parameter space.
The normal error is that ECLIPSE throws up negative compressibilities for your fluid.
If this happens then be sure to check your black oil tables covered the appropriate
range of pressures.
For a thorough description of the black oil model used by PVTi and how the tables are
extended above the bubble point see "Blackoil model" on page 372.
Can PVTi be used if you know the composition of

a fluid but do not have any observations? And
vice-versa?
You know the fluid composition
If you have compositional information about the fluid, but no observations, then you
can do anything you want within PVTi except use the Regression facility. This is
because the Regression facility tries to minimize the differences between lab
observations and PVTi’s theoretical predictions and this is not a sensible operation if
no observations are defined.
In fact, in PVT laboratories engineers tend to use the default fluid model (one that has
not been regressed on) for a given equation of state in PVTi to give them a ‘ball-park’
answer for their particular experimental observation. This can give them some idea as
to how to set their experiment up as they now have information on the kind of answers
they might expect to measure.
You have observations

If you have observations but no compositional information at all then PVTi does have
the facility to convert black oil tables into a fully compositional model. To do this you
must have black oil tables that were exported by PVTi using the 2002A release or later.
The tutorial entitled "Converting a black oil run to compositional" on page 65 will
explain this workflow in more detail.
Note If you have black oil tables exported from before and including 2001A_2 PVTi,
or you have no black oil tables at all then, unfortunately, PVTi will not be able
to construct the compositional model for you. !

Introduction
What black oil correlations are available in PVTi?
PVTi is a compositional PVT program and at the moment does not support black oil
correlations within its functionality.
It can however export a black oil model, using the compositional model tuned by the
user within PVTi to a DL or CVD experiment, for the ECLIPSE BlackOil simulator by
generating tables of Rs, FVF, etc. as a function of pressure for a given reservoir
temperature. To export a black-oil model select File | Export Keywords... . If you want
to export a black oil table from a DL experiment then you generally use the Oil
Reservoir... menu option and if you have a CVD experiment then the Gas
Reservoir...option is appropriate. For more information on exporting keywords see the
section entitled "Exporting keywords" on page 144.
How do I generate the asphaltene phase envelope

using PVTi?
Just like a phase curve has single phase regions for the vapor and liquid and a 2-phase
region an asphaltene phase envelope may exist for your fluid. The asphaltene phase
envelope partitions off a region in pressure-temperature space where an asphaltene
phase exists, analogous to the two-phase region in a standard phase curve.
The upper line partitioning a region in pressure-temperature space where an
asphaltene phase does and does not exist is called the Asphaltene Disappearance
Pressure (WDP) line. The lower line is called the Asphaltene Appearance Pressure
(WAP) line. In PVTi 2003A both curves are supported, whereas in PVTi
2002A/2002A_1 just the functionality for the appearance line is available.
Unlike a standard phase curve though where, as long as you do not have a fluid
consisting of a single pure component, there will always be a two-phase region, this is
not the case in terms of an asphaltene phase envelope. Asphaltene may never be
present no matter what temperature and pressure your fluid is at.
In order to try to generate one of the asphaltene curves a PNA split must first be
performed on the heaviest component/pseudo component. To do this select Edit | Fluid
Model | Split | PNA Distribution..., which does this split for you using correlations from
within PVTi.
Now, to create the curve an asphaltene appearance/disappearance experiment must
be performed. This can be done using the instructions in this section on "How do I
create an experiment along with a series of observations?" on page 17. In the
experiment entry panel tick the appropriate box depending on whether you want the
asphaltene disappearance/appearance curve and enter the temperature values where
you want the curve generated.
Caution You cannot plot theoretical predictions from PVTi of an experimental

unless at least one observation has been defined.
Make sure there is at least one observation defined for this experiment and then right
click on the experiment and select plot. You can super-impose your curve onto a
standard phase plot by using the super-impose button on the toolbar.

Introduction
Hint There is a shortcut for doing standard phase plots. This is accessed by clicking
and holding the left mouse button down on the appropriate fluid sample icon
and then dragging the cursor into the plot viewer.
There is a more detailed section on the theory behind how PVTi models the presence
of wax and asphaltene phases in "Wax and asphaltene precipitation in PVTi" on page
416.
How does PVTi support ECLIPSE Thermal?

The 2004A release of PVTi contains a new module that is capable of exporting a series
of keywords in a file suitable for use in an ECLIPSE thermal simulation. Therefore, just
as you were previously able to export PVT files suitable for use in ECLIPSE BlackOil
and ECLIPSE Compositional they can now do the same for ECLIPSE Thermal.
In PVTi and ECLIPSE Compositional we deal with an isothermal flash. This means
that, for a particular cell in a simulation, we know the composition of the fluid and the
pressure and temperature. We try to minimize the Gibbs Free Energy in order to
determine how each component splits across the different phases present. In other
words we try to find the K-values, which are the unknown variables.
In fact, in ECLIPSE Compositional, the temperature of each cell in a particular PVT
region is assumed not to change over time. In ECLIPSE Thermal this is not the case.
The temperature is a free parameter, which needs to be determined by way of a
different type of flash - one at constant energy. What we are saying here is that we
know the pressure, functional form of the K-values (K=K(P,T)), and the total energy of
the system and the temperature is unknown.
The new ECLIPSE Thermal Export facility allows you to export one of two keywords
(KVCR or KVWI) to model the functional form of the K-values for each component.
Since, as just explained, isothermal flashes are not performed in ECLIPSE Thermal,
PVTi also provides support in this export facility in terms of calculating density and
enthalpies in the oil and gas phases. In ECLIPSE Thermal the keywords PREFT, TREFT,
DREFT and ZFACTOR are used to calculate oil/gas densities and the keywords
STCOND, SPECHA, SPECHB, SPECHG, SPECHH and HEATVAPS are used to calculate oil
and gas phase enthalpies. For a brief summary of all the keywords exported for use in
ECLIPSE Thermal see "Outline of keywords for ECLIPSE Thermal" on page 387.
For more detailed information on the new thermal module see "ECLIPSE Thermal
Export Module" on page 423. For workflow information and a summary of the
exported keywords see "Compositional Data for ECLIPSE Thermal" on page 387.
How do I Use PVTi’s Batch Mode?

The batch mode facility is accessed from the command line (if using a PC) using the
command ‘$pvti -batch filename.pvi’ where filename is the rootname of your
project. If using a UNIX machine then use the command ‘@pvti -batch
filename.pvi’. The recommended way to prepare a file for use in the batch mode is
as follows:
1 Open the project file in interactive mode which you wish to use.
2 Save the file using the File | Save (concise)... menu option.

Introduction
3 Load this new ‘concise’ file back into PVTi and open the Set PVTi Program Options
panel by doing Utilities | Program | Options.... Set the last option Write Keywords
for Batch Mode to Yes.
4 Now perform the required workflow in interactive mode that you wish the batch
mode to reproduce e.g. simulation of experiment, splitting, grouping, export, etc.
5 When you have performed the workflow save the file using File | Save...
6 The saved file is now suitable for running in batch mode.
For a comprehensive review of the new batch mode functionality see "Batch system
and keywords" on page 163.
How Can I Export PVT Tables to use the API

Tracking Functionality in ECLIPSE BlackOil?
PVTi now has the capability to export multiple PVT Tables for use in ECLIPSE
BlackOil’s API Tracking. Without the API Tracking facility, the presence of different
types of oil in the reservoir could be handled with the aid of PVT region numbers. Oil
in PVT region 1 would have its properties determined from PVT table number 1, and
so on. However, this method cannot model the mixing of oil types. Oil flowing from
region 1 into region 2 would appear to take on the properties associated with region 2
Just like exporting a standard black oil table a depletion experiment must be defined
in order to do API Tracking export. The workflow is as follows:
1 Use the File | Export Keywords | API Tracking option in ECLIPSE BlackOil... menu
option to open the Export Panel for API Tracking.
2 Select the samples you wish to use in the export and move them to the Use box.
3 Choose the keywords you wish to export. Normally the Live Oil (PVTO) and Dead
Gas (PVDG) keywords are used with the API Tracking option.
4 Select whether you wish to export a gas table for each sample or just a single gas
table.
5 Select whether you wish to plot the tables in PVTi.
6 Select whether you wish to write to tables using full double precision numbers.
7 Select the separator you wish to use for the export.
8 Select the units you wish to export the table in.
PVTi should then write the tables to a file and show them in the output display. This
file is then suitable to use as the PVT input for an API Tracking run in ECLIPSE
BlackOil. For a similar description of the API Tracking workflow see "Export for API
Tracking option in ECLIPSE BlackOil" on page 145. For a technical description of the
API Tracking model in ECLIPSE as well as an explanation of how PVTi calculates
suitable PVT tables see "Model for API Tracking option in ECLIPSE BlackOil" on page
383.

Introduction
Introduction
Chapter 3
General information
The PVTi program is an Equation of State based package for generating PVT data from
the laboratory analysis of oil and gas samples. The program may be used through an
interactive menu system or run in a batch mode. An interactive session can be saved as
a batch input file, which contains commands to reproduce the interactive operations.
Alternatively, a batch input file can be run from an interactive session.
Equations of state and viscosity correlation

Four equations of state are available, implemented through Martin’s generalized
equation. This enables the Redlich-Kwong, Soave-Redlich-Kwong, Peng-Robinson
and Zudkevitch-Joffe equations to be used. Two 3-parameter extensions of the Peng-
Robinson Equation of State are also available, one based on a Peneloux et al. volume
shift, the other being an implementation of the Schmidt-Wenzel Equation of State 2-
parameter Peng-Robinson. The Soave-Redlich-Kwong Equation of State similarly has
a three-parameter extension.
Viscosities may be calculated using a method by Pedersen et al. based upon a
corresponding states comparison with methane, or by the Lohrenz-Bray-Clark
method.
Fluid definition
Multiple fluid samples can be defined by specifying components as one of three types.
Library components require only that the appropriate component mnemonic be
entered. Characterized components define properties of plus fractions from a limited
set of information. Finally all the properties of a component can be defined, a facility
which can be used selectively to edit the properties of existing components.
Introduction 27
General information
It is possible to group the components to reduce or pseudoise the fluid system (make
a fluid definition of the system using pseudo components), or to split the plus fraction
into more components, preserving molecular weight and mole fraction. Multiple
samples having different plus fraction properties, say mole weight and specific gravity,
can be characterized by splitting the plus fraction into two or more pseudo-
components of fixed properties but variable composition.
Fingerprint plots of mole fraction against molecular weight, or phase diagrams, are
available.
Material balance checks

A compositional material balance can be performed on any gas condensate or volatile
oil for which a laboratory constant volume depletion or differential liberation
experiment has been performed. This can be used to estimate liquid compositions and
hence K-values. The calculated quantities can then be used to estimate the quality and
consistency of the laboratory data. Additionally, tests on recombination of separator
data can be performed and estimates of reservoir recovery can be made.
Simulation of experiments
Experiments may be performed on the fluid systems defined using the equation of
state model. Possibilities are:
• saturation pressures
• flash calculations
• constant composition expansions
• constant volume depletions
• differential liberations
• swelling tests
• multi-stage separator simulations.
Other experiments available are:
• composition versus depth
• vaporization test
• multiphase flash
• critical point
• saturation temperature
• first contact miscibility
• multiple contact miscibility (condensing and vaporizing).
• wax appearance temperature
• asphaltene appearance pressure
28 Introduction
General information
Regression of equation-of-state to measured data
The equation of state model may be tuned by regression. The critical state data, Ω a and
Ω b values, interaction coefficients, δ ij , and volume shift parameters, s i (for the three-
parameter volume shift equations of state), may be matched to experimental data from
the first eight of the experiments mentioned in the previous paragraph. Additionally,
depending on the use of certain facilities and options, five special regression
parameters are also available. These are the A coefficient in the Cheuh-Prausnitz Bids,
the thermal expansion coefficient in the modified Peneloux et al. volume shift method,
and three variables associated with the Modified Whitson splitting technique (denoted
Semi-Continuous-Thermodynamics), being the mole weight and distribution
skewness parameter (on a sample-by-sample basis) and the characterization factor of
the plus fraction.
Hint Almost any result from the experiments mentioned can be used as an
observation against which to regress.
PVT data for ECLIPSE simulators

Black oil and equilibration tables for ECLIPSE can be prepared, using the liquid and
gas compositions obtained from constant volume depletion or differential liberation
experiments passed through a separator system using the Coats or Whitson and Torp
methods.
Gas injection schemes can be modeled using the ECLIPSE pseudo-compositional
model which modifies the ECLIPSE black oil tables as a function of the volume of
injected gas.
For use in ECLIPSE Compositional, either pseudoised/regressed equation of state data
or tabular data (either K-values or liquid and vapor mole fractions as a function of
pressure) can be output.
Black oil tables can also be generated for the Vertical Flow Performance (VFPi)
program, simulating the phase and volumetric change in the wellstream fluid by a
constant composition expansion experiment at two temperatures, the highest
(reservoir) and lowest in the production string.
Water properties may also be output for use in any of the above programs. Default
values for formation volume factor, compressibility, etc., are constructed using well-
known correlations from specification of the pressure, temperature, salt and gas
content of the water, though these may be changed.
Introduction 29
General information
30 Introduction
General information
Getting started
Chapter 4
Starting PVTi
Windows NT and Windows 98 platforms

GeoQuest Program Launcher
1 Start the GeoQuest Program Launcher.
2 Click the PVTi button.
3 Select the version and working directories as required.
Command line
1 Type the command $PVTI in a command prompt window.
UNIX platforms
1 Type the command @pvti at the command prompt.
Getting started 31
Starting PVTi
32 Getting started
Starting PVTi
Tutorials
Chapter 5
Overview
The tutorials provide a step-by-step introduction to the functionality of PVTi.
Note These tutorials are not intended to teach PVT analysis, but instead concentrate
on illustrating typical work-flows for PVTi.
Each tutorial is divided into a number of distinct sections intended to highlight a

specific aspect of the analysis process. To avoid repetition, later tutorials assume
familiarity with some used in the first tutorials, so it is strongly recommended that you
work through them in the order they are presented.
Available tutorials
The following tutorials are available:
• "Fluid Properties Estimation" on page 35
• "Creating a fluid system" on page 39.
• "Simulating experiments" on page 45.
• "Fitting an equation of state to experimental results" on page 53.
• "Exporting ECLIPSE Black Oil PVT tables" on page 59.
• "Converting a black oil run to compositional" on page 65.
• "Workflow Tutorial" on page 69.
• "Multiphase Flash" on page 77.
• "Exporting an ECLIPSE Thermal model" on page 81.
• "Data analysis and quality control" on page 85.
• "Removing contamination from samples" on page 93.
Tutorials 33
Available tutorials
• "Converting old projects to the current version" on page 97.
34 Tutorials
Available tutorials
Fluid Properties Estimation
This tutorial describes how to use PVTi for Fluid Properties Estimation.
The data for this tutorial is provided with the standard installation of PVTi in the
directory:
$ECL/2005a/pvti/tutorials
and you should copy the files from this directory to your local directory before starting
the tutorial.
The tutorial is split into several sections:
• "Introduction" on page 35
• "Basic Information - Fundamentals" on page 35
• "Flash calculations" on page 37
• "Discussion" on page 37
Introduction
Fluid properties estimation can provide quick-look PVT tables at the well site.
A saturation pressure (bubble or dew-point) together with a reservoir composition are
sufficient inputs to provide a quick-look simulation, giving an initial estimation of
fluid properties in advance of a full fluid analysis in the lab.
After completing this tutorial you should be able to use PVTi as a simulation tool for
fluid properties estimation.
Basic Information - Fundamentals

1 Start PVTi (if you are unsure about this see "Starting PVTi" on page 31).
2 Enter FPE.PVI as the file name for the new project.
Hint When a new, empty project is created in PVTi, the Fundamentals panel opens
automatically. To re-enter this panel at any later time, select PVTi: Edit |
Fundamentals...
The Fundamentals panel allows you to enter the minimum information required
to create a complete equation of state model.
3 Click on the Enter Weight Fractions check box.
Hint The mole fractions that you see in lab reports are derived from weight fractions
and the mole weights of the components. It is weight fractions that are actually
measured. PVTi allows a choice of either form in the Fundamentals panel.
Tutorials 35
Tutorial 1: Fluid Properties Estimation
4 Right-click in the table and select Table Import | From File.
• Import the file FUNDAMENTALS.TXT .
• In the Text Import Wizard switch on Ignore Records and set the number of
records to ignore to 2 (since we want to ignore the column headings).
The Fundamentals panel should now look like Table 5.1.
Table 5.1 The Fundamentals panel
Components ZI Weight Frac. Mol Weight Spec Gravity

(percent) (percent)
CO2 0.43
N2 0.05
C1 6.25
C2 3.10
C3 3.27
IC4 0.89
NC4 2.44
IC5 1.11
NC5 1.09
C6 3.88
C7+ 77.49 218
Hint Only enter mole weights for components whose properties will be
characterized, the other components come from the library. Also, specific
gravity is an optional additional parameter that can be used in the
characterization, if it is not specified it is calculated using a correlation.
5 Click Apply
PVTi adds the mole fractions and the specific gravity of the plus fraction.
6 Click on the Fluid Properties Estimation (FPE) check box
The temperature and pressure fields are now active.
• Enter a Temperature of 220 F.
This is the temperature of the saturation pressure (dew or bubble point) and
the temperature that is used in the generated depletion experiments.
• Enter a Saturation Pressure of 2800 psia.
Hint PVTi uses this saturation pressure to fit the fluid model. The weight of the
plus-fraction is varied, the weight fraction being constant, until the saturation
pressure predicted by the equation of state matches the entered pressure.
• Enter a Maximum Pressure of 5000 psia.
36 Tutorials
Hint This is the maximum pressure in the depletion experiments that are created.
7 Set the Project Units to Field, this sets the units that are used on the plots.
8 Click OK.
This is all the information required to fit the equation of state and to generate the
Constant Composition Expansion (CCE), depletion experiment (either differential
liberation or constant volume depletion) and the optimized separator. After fitting the
equation-of-state and creating the experiments, default plots from the depletion
experiments are drawn along with the phase curve for the fitted fluid.
The methods used in Fluid Properties Estimation are explained in "Fluid Properties
Estimation" on page 406.
9 PVTi: Run | Simulate
This opens a complete report for the project including the results from all the
created experiments.
Hint By clicking on one of the experiments in the sample tree with the right mouse
button, and selecting Report..., you can view the reports for individual
experiments separately.
Flash calculations
1 Right-click on ZI in the project tree-view and select Properties Estimation (FPE)...
• Enter a temperature of 60 F
• Enter a pressure of 15 psia
• Click OK.
2 Right-click on the newly created Flash simulation (FLASH1) and select Report to
see the results of flashing the reservoir fluid to standard conditions.
This allows you to attempt any Flash calculation on the reservoir fluid.
Hint The Properties Estimation panel can also be used to create additional
separators, saturation pressure or depletion experiments, for example at other
temperatures.
Discussion
Fluid Properties Estimation is a useful tool, particularly in situations where full lab
analysis of the fluid is not available for a complete equation-of-state matching project.
For full details of the operations performed during fluid property estimation see "Fluid
Properties Estimation" on page 406.
Tutorials 37
During fluid properties estimation, the project created is a complete PVTi project. This
means that a more experienced user has access to the rich range of features within the
product. At the same time, the less experienced user can use PVTi for Fluid Properties
Estimation without requiring in-depth knowledge of equation-of-state methods or
PVT analysis.
38 Tutorials
Creating a fluid system
This tutorial describes how to specify fluid properties in PVTi. It covers the basic
functionality of PVTi; knowledge of this tutorial is assumed in the later tutorials, so you
are advised to work through them in order.
The data for this tutorial is provided with the standard installation of PVTi in the
directory:
and you should copy the files from this directory to your local directory before starting
the tutorial.
• "Defining a fluid" on page 40
• "Selecting an equation of state" on page 41
• "Program options" on page 41
• "View fluid attributes" on page 42
• "Saving the SYSTEM section for future use" on page 43
Introduction
The PVT report for this fluid contains details of three experiments: a Constant
Composition Expansion experiment, a Differential Liberation experiment, and a
Bubble Point experiment. The later tutorials describe how the experimental results
may be used to fit an equation of state to the experimental behavior, and how this fitted
equation of state can be used to generate PVT tables for use in reservoir simulations.
This tutorial shows how to set up basic fluid properties in PVTi and how to display the
phase envelope for the defined fluid.
Tutorials 39
Tutorial 2: Creating a fluid system
Defining a fluid
PVT analysis involves fitting an Equation of State to experimental data and then using
this Equation of State to produce PVT tables for use in reservoir simulations. The first
step is to start up PVTi, and import the component and fluid definitions.
Table 5.2 shows the component and fluid definitions that are used in this tutorial.
Table 5.2 Component and fluid definitions
Component % Mole Fraction Mole Weight Specific Gravity

CO2 0.91
N2 0.16
C1 36.47
C2 9.67
C3 6.95
IC4 1.44
NC4 3.93
IC5 1.44
NC5 1.41
C6 4.33
C7+ 33.29 218.0 0.8515
2 Select PVTi: File | New...
3 Enter BLACK.PVI as the project name in the file selection window.
4 Click on Open on PC or OK on UNIX platforms.
Note This has defined BLACK as the prefix for any files that are written by PVTi.
The Fundamentals panel opens so that basic project information can be entered.
5 Enter CO2, N2, C1 and C6 into the Components column.
6 Click Apply.
7 Click Yes so that PVTi fills in the library component names.
8 Enter the mole fractions from Table 5.2 and the details for the C7+ component into
the Fundamentals panel and click OK.
Note The components for which no mole weight or specific gravity has been
specified are automatically set to use the PVTi component properties library
(see "Component types" on page 113.)
40 Tutorials
Hint The component properties can be examined by selecting PVTi: Edit | Fluid
Model | Components.... This panel can also be used to add additional
components, the select alternative characterization methods and to manually
defined component properties.
Hint Only one sample, ZI, is defined in the Fundamentals panel. Additional
samples can be created using PVTi: Edit | Samples | Names... and mole
fractions can be entered using PVTi: Edit | Samples | Compositions...
Selecting an equation of state

In this tutorial, the three-parameter Soave-Redlich-Kwong equation of state (see
"Equation of state formulation" on page 331) is fitted to the results of experiments
carried out on the fluid defined in "Defining a fluid" on page 40. The Lohrenz-Bray-
Clark correlations (see "Lohrenz, Bray and Clark" on page 345) is used for viscosity.
1 PVTi: Edit | Fluid Model | Equation Of State...
This opens the Equation of State and Viscosity Correlation panel.
2 Select the 3-parameter Soave-Redlich-Kwong equation of state (SRK3).
3 Click on OK.
4 Click on OK to change the parameters to SRK3 defaults.
Program options
1 PVTi: Utilities | Program | Options...
This opens the Program Options panel.
2 Set the Separator GOR calculation to use Liquid at Stock Tank Conditions.
3 Set the Temperature-dependence for volume shifts to be calculated by Polynomial
correlations. (See "Shift parameters" on page 336.)
The Program Options panel should now contain the following data:
Table 5.3 Program Options data
Field Data
Definition of Liquid Saturation in CCE Sliq=Vliq/Vsat
Treatment of volume shifts Dependent
Separator GOR Calculation Liquid at Stock Tank Conditions
Temperature dependence for volume shifts Polynomial correlations
Specify/Calculate density and molar volume units User units
Specific Heat Capacity Coefficients and Calorific
Values
Calculated compositions No Save to samples
Component Library Katz-Firoozabadi
Tutorials 41
Table 5.3 Program Options data (Continued)
Field Data
Experimental Compositions Output to Screen/PVP
Experimental Results Always Output to PVP
Plot Vectors No Output to file
Print File Output A4 format
Definition of GOR in Diff Lib Normal
Definition of Oil relative volume in Diff Lib Oil FVF = Voil(p)/Voil(stc)
Black Oil Table Output All Data
Flash Calculations E300 Flash
Sample mole fractions when regressing Keep Fixed
4 Set Treatment of Volume Shifts to Independent and click on OK.
View fluid attributes

Now that a fluid has been defined, there are two plots available to review the fluid we
have entered. First is the fingerprint plot of mole fraction versus molecular weight; the
second is a phase plot.
1 Right-click on ZI in the project tree-view and select Fingerprint Plot from the pop-
up menu.
The plot should look like Figure 5.1.
Figure 5.1 Fingerprint Plot
2 PVTi: View | Samples | Phase Plot...
42 Tutorials
3 Request Sample ZI, 5 quality lines.
4 Click on OK.
The plot should look like Figure 5.2.
Figure 5.2 Phase Plot
Note A default phase plot with a single quality line can be generated by dragging ZI
from the tree-view of the project (in the left pane of the main window), and
dropping it on to the main plot workspace.
Saving the SYSTEM section for future use

The fluid sample definition can output as the RUNSPEC and SYSTEM sections of a PVI
file.
1 PVTi: File | Save (Concise)...
2 Call the file FLUID_DEF.PVI.
Hint The complete project can be saved using PVTi: File | Save... This, effectively,
saves a history of the project. The original fluid description is saved, along
with SPLIT or GROUP sections for split and group operations you perform. By
choosing to save current modifications, the system is saved in its current state,
after all splits, groups, etc., have been performed. For work in progress it is
usually better to use Save so that past steps can be recovered. For a final fluid
model, the Save (Concise) option allows a complete description of the final
model to be saved, without the steps taken to get there.
This file can now be read in using the PVTi: File | Import | SYSTEM... option.
Tutorials 43
Discussion
In this tutorial a fluid was defined from data and an Equation of State was selected to
describe that fluid. This fluid definition provides the basic building blocks for further
PVT analysis. The fluid definition can be used in simulation studies to compare it with
experimental results from the reservoir fluid; see "Simulating experiments" on page 45.
This definition can then be adjusted so that it describes the experimental results; see
"Fitting an equation of state to experimental results" on page 53. The fitted fluid
definition is finally used to generate PVT tables for ECLIPSE ; see "Exporting ECLIPSE
Black Oil PVT tables" on page 59, ECLIPSE pseudo-compositional, VFPi and ECLIPSE
Compositional.
44 Tutorials
Simulating experiments
This tutorial illustrates how experiments are simulated in PVTi. It covers the basic
functionality of PVTi. Knowledge of this tutorial is assumed in later tutorials, so you
are advised to work through them in order. The data for this project are provided in the
standard installation of PVTi in the directory:
and should be copied to the local directory before starting the tutorial.
• "Defining experiments for simulation" on page 46
• "Plotting simulation results" on page 48
• "Defining further experiments and observations" on page 48
• "Defining further experiments and observations" on page 48
• "Simulating all the experiments" on page 50
• "Discussion" on page 51.
Introduction
This tutorial describes how experimental observations can be entered into PVTi and
how the experiments can then be simulated from an existing fluid definition.
2 File | Open...
3 Open the file FLUID_DEF.PVI that was created in the last tutorial. (Alternatively
open the supplied tutorial file FLUID_CORRECT.PVI).
Setting units
1 Utilities | Units...
2 Set the Unit Type to Field
3 Set the Temperature Unit Type to Fahrenheit
4 Set Mole Fraction or Percentage to Percentage
5 Set Absolute or Gauge Pressure to Gauge.
6 Click on OK.
Tutorials 45
Tutorial 3: Simulating experiments
Defining experiments for simulation
In this part of the tutorial, the experimental results from the PVT report are brought
into PVTi ready for simulation. In this section data from a constant composition
expansion experiment are brought into PVTi. If you do not have access to a
spreadsheet, type the numbers from the tables into the data entry forms in PVTi at the
appropriate points.
Table 5.4 Constant Composition Expansion experiment at 220o F (* indicates

bubble point pressure)
Pressure Relative Volume

(PSIG) (V(p)/V(pb)
5000.0 0.9453
4500.0 0.9541
4000.0 0.9638
3500.0 0.9746
3000.0 0.9867
2900.0 0.9893
2800.0 0.9920
2700.0 0.9948
2620 0.9970
2605 0.9974
2591 0.9978
2516.7* 1.0000
2401 1.0243
2253 1.0599
2090 1.1066
1897 1.1750
1698 1.2655
1477 1.4006
1292 1.5557
1040 1.8696
830 2.2956
640 2.9457
472 3.9877
1 PVTi: Edit | Experiments...
2 Experiment Entry: Add | Pressure Depletion | Constant Composition Expansion...
46 Tutorials
Hint The constant composition expansion or CCE experiment can sometimes be
known as a constant mass study in PVT Reports.
The differential liberation or DL experiment is also known as a differential

vaporization experiment in PVT Reports.
The multi-stage separator or SEPS experiments can also be called a separation

test in PVT Reports.
The Experiment Entry window now shows three folders: General, Observations
and Components. These folders are used to define the experiment entry form.
3 Select the Observations folder.
4 Click in the top left cell of the table and select Pressure from the drop-down list in
that cell.
5 In the second column select Relative Vol. from the drop-down list.
Hint By making the column headings the same as those in Table 5.4, the task of data
entry is simplified. The ability to tailor the table means that data entry can then
be further accelerated by importing observations from a text file or the
clipboard.
6 Click on Next.
The table now shows two folders. The Components folder has disappeared as
there were no component observations selected; the General folder now shows an
entry field to select fluid type and another to enter the temperature of the
experiment.
7 In the General folder, enter the temperature from Table 5.4 (220 F).
The Observations folder now shows a two-column table with the columns selected
previously. The table resembles Table 5.4.
Table 5.4 is provided in the file CCE_TABLE.TXT
9 Right-click in the table and select Table Import | From file...
10 Select CCE_TABLE.TXT and click on Open.
11 In the Text Import Wizard turn on Ignore Records and set the number of records to
ignore to 1 (since we want to ignore the column headings).
The view of the table should no longer contain the first row.
12 Click on OK.
Note The error message “Cannot delete rows from this table” appears
This is because the table has a fixed length and the file we are importing from
has fewer rows than the table. This message can be safely ignored.
13 Click on OK to remove the message “Cannot delete rows from this

table”.
Tutorials 47
The table now contains the same information as Table 5.4. As the experiment
information is complete, the experiment can be created.
14 Click on Next to create the experiment.
Hint The data tree now shows the created experiment (CCE1). The asterisk (*) next
to the experiment’s name means that it is active. CCE1 has one observation
node, for the relative volume measurements.
15 Click Close to shut the panel.
Plotting simulation results

1 Click on the Relative Vol. observation in the Data Tree and drop it over the Main
Plot Window.
The Main Plot Window should now look like Figure 5.3.
Figure 5.3 The plotted simulation results
Defining further experiments and observations

In this section of the tutorial the other experiments are defined, along with their
observations. The equation of state is later fitted to these observations, and then the
fitted equation is used to generate tables for a fully compositional ECLIPSE simulation.
48 Tutorials
Differential liberation experiment
The first experiment to be added is a differential liberation experiment (Table 5.5), as in
"Defining experiments for simulation" on page 46.
Table 5.5 Differential Liberation Experiment at 220o F (* indicates bubble point

pressure)
Gas Reservoir Gas

Oil Solution Oil Gas Volume
Pressure Deviation
Volume GOR Density Relative Factor
(PSIG) Factor
Factor (Mscf/stb) Density
Z (lb/ft3) (rb/Mscf)
2516.7* 1.7493 1.1342 45.110
2350 1.7095 1.0526 0.8686 45.6688 0.7553 1.2574
2100 1.6535 0.9378 0.8692 46.5022 0.7547 1.4070
1850 1.6013 0.8309 0.8719 47.3311 0.7565 1.6006
1600 1.5523 0.7307 0.8767 48.1595 0.7614 1.8586
1350 1.5057 0.6361 0.8836 48.9920 0.7704 2.2164
1100 1.4609 0.5460 0.8926 49.835 0.7859 2.7411
850 1.4171 0.4591 0.9036 50.6992 0.8121 3.5773
600 1.3726 0.3732 0.9167 51.6076 0.8597 5.1050
350 1.3234 0.2824 0.9324 52.6319 0.9618 8.7518
159 1.2720 0.1960 0.9481 53.6731 1.1726 18.6846
0 1.1228 0.0 56.3229 1.8901

2 Experiment Entry: Add | Pressure depletion | Differential Liberation...
3 In the Observations folder, set the table headings to match those in Table 5.5:
Pressure, Oil Rel. Vol., Gas-Oil ratio, Vapor Z-factor, Liquid Density, Gas Gravity,
Gas FVF.
4 Click on Next
5 Enter 220 F as the temperature in the General folder.
The file DL_TABLE.TXT provides the table in Table 5.5.
6 Import the file DL_TABLE.TXT into the table in the Observations folder,
remembering to ignore the first line, which contains column headings.
The Experiment Entry panel now shows that there are 2 experiments defined.
Defining the bubble point experiment

Finally, there is a bubble point experiment at 220o F to be added.
8 Experiment Entry: Add | Single Point | Bubble Point...
9 In the Observations folder set the first column heading to Sat. Pressure and the
second to Liquid Density
Tutorials 49
10 Click on Next
11 Enter the temperature, 220o F in the General folder.
13 Enter the saturation pressure as 2516.7 psig and the liquid density as 45.11
lb/ft3.
15 Click Close.
Simulating all the experiments

All the experiments have now been entered. In summary, then, the project should now
contain the following:
• A fluid description (component properties and a sample defined by mole
fractions of components).
• A Constant Composition Expansion experiment with observations of relative
volume.
• A Differential Liberation experiment with observations of: relative oil volume,
solution gas-oil ratio, Z-factors, oil density, gas gravity and gas formation
volume factor.
• A Bubble Point experiment at 220o F with observations of bubble point
pressure and liquid density.
Hint The information about which experiments have been defined, and for which
observations have been entered for those experiments, is contained in the Data
Tree.

A simulation report, showing information on all the experiments, is displayed in
the Output Display panel.
2 Output Display: File | Close
Plotting all observations for an experiment

3 PVTi: View | Observations...
4 Select the Differential Liberation (DL1) experiment.
5 Click OK.
This plots each observed data set (as points) for the differential liberation
experiment and each calculated data set (as lines) generated by simulation.
Hint Double-clicking on one of the small plots swaps it with the large plot.
6 Examine each of the plots and note how well (or badly) the simulation has matched
the data.
50 Tutorials
Hint You can right-click on an axis and select Show Edit Box from the drop-down
menu. This opens the Axis Property Editing panel. In this panel you can
customize the axes, for example by changing the units used in plotting.
In the next section, the match between calculated and observed data values are
improved by regression.
Saving the project for future use

The fluid sample definition can be output as the RUNSPEC, SYSTEM and SIMULATE
sections of a PVI file.
1 PVTi: File | Save (concise)...
2 Call the file SIMULATE_SECTION.PVI.
Discussion
In this tutorial an existing project was extended to include experiments. Constant
Composition Expansion, Bubble Point and Differential Liberation experiments were
imported and simulated for the defined fluid. The match between the experimental
observations and the simulated results was examined using the plotting facilities in
PVTi. The fluid model can then be adjusted so that it provides the best fit (in a least-
squares sense) to the experimental observations (see "Fitting an equation of state to
experimental results" on page 53). The fitted fluid definition is finally used to generate
PVT tables for ECLIPSE (see "Exporting ECLIPSE Black Oil PVT tables" on page 59).
Tutorials 51
52 Tutorials
Fitting an equation of state to experimental results
This tutorial shows how a fluid definition can be fitted, by regression, to describe
experimental results. This tutorial covers basic functionality of PVTi and knowledge
this tutorial is assumed in later tutorials, so you are advised to work through them in
order. The data for this project are provided in the standard installation of PVTi under
the directory:
• "Fitting an equation of state by regression" on page 53
Introduction
This tutorial illustrates the fitting of the fluid definition to the experimental
observations.
The fluid definition and experiments are read in from an existing PVI file and the
regression facilities in PVTi are used to generate an improved fit between the two.
Fitting an equation of state by regression

2 File | Open...
3 Open the file SIMULATE_SECTION.PVI created in the last tutorial. (Alternatively
you can use the file SIMULATE_CORRECT.PVI.)
Hint The Data Tree shows the contents of the opened project.
Fitting an equation of state by regression

In this part of the tutorial, the equation of state is fitted to the observation data to
produce a better representation of the fluid. A sensitivity analysis is carried out to
determine which attributes of the fluid components improve the solution by the
smallest change. The most sensitive attributes are then adjusted slightly by regression
to improve the equation of state model of the fluid.
Tutorials 53
Tutorial 4: Fitting an equation of state to experimental results
The first step in designing any regression is to determine the parameter set that
will be used. There are certain steps an engineer can take to improve the
performance of a regression. The first step is to try to make all regression variables
have similar sizes. This is done to prevent a minor constituent of the fluid having
its properties varied extensively to achieve a mathematical fit, which is not a
reasonable physical solution.
4 Examine the fluid component data in "Component and fluid definitions" on page
40.
Note The idea here is to look for consecutive components that have small mole
fractions. These can be grouped together and treated as a single regression
variable, forcing the solution to be physically realistic.
Hint The properties of C1 and C2 are well known and generally do not differ
significantly from the library properties.
Grouping the C3, IC4, NC4, IC5, NC5 and C6 components into a single regression
variable preserves monotonicity between the components, in addition to creating
a variable that accounts for 19.5% of the total composition.
Hint The plus fraction (C7+) contains a mixture of components C7+ and higher, so
its properties may not be so well-determined. This makes the plus fraction an
ideal candidate for regression to fit the equation of state to the experimental
results.
C7+ is the second regression variable.
Sensitivity analysis
Sensitivity analysis is used to establish which fluid properties most affect the difference
between the observed and simulated values. The sensitivities are calculated for critical
temperature and pressure for each experiment, for both regression variables. Finally
the most sensitive properties will be selected for use in the regression.
Hint In any regression, having a few very sensitive parameters is preferable to

having hundreds of insensitive ones. Always look for parameters that can be
discarded (this is called conditioning the problem - an ill-conditioned problem
is difficult to solve).
5 PVTi: Run | Regression... opens the Regression panel.

6 Select Normal as the Type of regression variables in the Variables section of the
panel.
7 Click Variables.
The regression variables are numbered for each property. Entering 1 in the critical
pressure (Pcrit) column in the rows corresponding to C3, IC4, NC4, IC5, NC5 and
C6 groups those components into the first regression variable.
54 Tutorials
8 Fill in the table in the Select EOS parameters for regression panel with the
following data:
Mnem Pcrit Tcrit

C3 1 1
IC4 1 1
NC4 1 1
IC5 1 1
NC5 1 1
C6 1 1
C7+ 2 2
9 Leave the second part of the Select EOS parameters for regression panel blank.
10 Click on OK.
Hint The second part of the Select EOS parameters for regression panel relates to
binary interaction coefficients.
11 Click Regression in the Report section of the panel

The Regression Report panel provides several views of the regression problem,
designed to give the best possible insight into creating a fluid model.
For a description of the Regression Report panel see "Regression Report" on page
142.
• Select the Sensitivities folder.
The sensitivities for the first Pcrit parameter are generally lower than for the
other regression variables.
• Select the Hessian folder.
The values in the leading diagonal dominate the matrix, except in the first row,
the row relating the first Pcrit parameter.
• Select the Covariance folder
In this table the largest value is for the first Pcrit parameter, indicating that it is
the least well determined by the regression.
• Select the Correlation folder.
There is a strong negative correlation between the two Pcrit parameters,
indicating that the regression would proceed better if only one of those two
parameters were used.
From an examination of the information in the Regression Report panel, it can be
seen that the first Pcrit parameter is not likely to aid the regression, and it may
hinder it. Consequently that regression variable is removed before regression is
started.
12 Close the Regression Report panel.
13 Click Variables in the Regression panel.
14 In the Select EOS parameters for regression panel click on Reset to clear all the
cells in the table.
Tutorials 55
15 Fill in the columns to describe the reduced set of regression variables with the
following data:
Mnem Pcrit Tcrit

C3 1
IC4 1
NC4 1
IC5 1
NC5 1
C6 1
C7+ 1 2
16 Click on OK.
Viewing the regression progress

The results of regression are viewed in a similar way to the results of simulation.
17 Right-click on experiment DL1 in the project tree-view and select Plot from the
pop-up menu.
18 Click Run in the Regress section of the Regression panel.
This starts the regression.
19 Click on Regression in the Report section of the Regression panel.
• Select the Modifiers folder.
The difference between the final and initial value of each regression variable is
displayed.
• Select the Details folder.
An observation-by-observation breakdown of the final fit is shown, along with
the total fit to all data (both unweighted and incorporating the observation
weights).
20 Examine the plots in the main window.
The observed data are plotted as points and the simulated data before and after
regression are plotted as lines.
The regression has improved the equation of state model, so the regression results
can be accepted.
Hint Right-clicking on an experiment allows you to choose to turn off that

experiment during the regression process.
21 Click Accept in the Regress section of the Regression Panel.

22 Close the Regression Report panel.
Saving the project

1 PVTi: Save (concise)...
56 Tutorials
2 Call the file REGRESS_SECTION.PVI
The results of regression are the fluid definition (that is the SYSTEM section) of the
newly created PVI file. They can now be read in using the PVTi: File | Import PVI
Section | SYSTEM... option.
Discussion
In this tutorial a fluid definition, an Equation of State and some experiments along with
their observations were imported from an existing PVI file (the file created in
"Simulating experiments" on page 45). The definition was adjusted so that it provided
the best fit (in a least-squares sense) to the experimental observations ("Fitting an
equation of state to experimental results" on page 53). This fitted fluid definition can
now be used to generate PVT tables for ECLIPSE (see "Exporting ECLIPSE Black Oil
PVT tables" on page 59).
Tutorials 57
58 Tutorials
Exporting ECLIPSE Black Oil PVT tables
This tutorial provides a typical workflow for PVTi: producing PVT tables that are then
used in an ECLIPSE BlackOil simulation. The data for this project are provided with
the standard installation of PVTi under the directory:
• "Exporting water properties" on page 59
• "Generating ECLIPSE Black Oil PVT tables" on page 60
• "Importing the keywords into ECLIPSE Office" on page 62
Introduction
Once the fluid description has been fitted to the experimental observations, it may be
used in a reservoir simulation. PVTi facilitates the transition between a fluid
description and the PVT keyword description required by the ECLIPSE family of
simulators.
In this tutorial PVT tables are created for the fluid definition developed in the tutorials
2, 3 and 4. The output tables are then used in an ECLIPSE simulation.
2 Open REGRESS_SECTION.PVI created in the last tutorial (alternatively, open
REGRESS_CORRECT.PVI).
Exporting water properties

The water properties exported from PVTi are generated by correlation. This is
effectively separate from the fluid model.
1 PVTi: File | Export Keywords | Water...
• Enter a reservoir temperature of 220 F and an initial reservoir pressure of 2500
psig.
• Click on OK
• Enter the filename PVTW.PVO for the water keyword
2 Close Output Display panel.
Tutorials 59
Tutorial 5: Exporting ECLIPSE Black Oil PVT tables
Generating ECLIPSE Black Oil PVT tables
In order to generate ECLIPSE BlackOil simulation PVT tables, PVTi requires either a
Differential Liberation experiment or a Constant Volume Depletion experiment to be
simulated from the fitted equation of state. The PVT tables are generated off either of
these experiments.
1 Right-click on experiment DL1 in the sample tree and select Export Keywords...
from the drop-down menu
Hint PVTi: File | Export Keywords | Oil reservoir... produces an export panel for all
available Differential Liberation experiments.
2 Select PVTO and PVDG (Live oil and dry gas) on the radio button menu.
The Separators drop-down menu becomes active. This is because the produced
fluid from the Differential Liberation experiment must be passed through a surface
separator to calculate, for example, surface gas-oil ratios. The default is a separator
at Standard Conditions. If any separator experiments were defined for this sample,
they would also appear here.
3 Click OK
4 In the File Selection box, enter ECLIPSE100 as the name of the export file.
The keywords are generated and the Display Output module shows the generated
keywords.
Note The comments prefixed with --PVTi that appear before each keyword are the
concise version of the current PVTi project. This is the minimum information
PVTi requires to create the tables. This information can be used to rapidly
convert an ECLIPSE BlackOil data-set to an ECLIPSE Compositional data-set.
Caution Avoid editing the --PVTi prefixed comments. Any changes may
invalidate the file, preventing PVTi from reading it.
Note The sample SWELLSAM has been added to the sample tree. This sample is the
swelled sample that was obtained by swelling the original sample with vapor
that was split off just below the bubble-point of the fluid. PVTi automatically
swells the sample with the vapor from the bubble-point so that the table can
be extended to values above the original bubble point.
The information in the keywords is also shown in the main plot space, keyword PVTO
is shown in Figure 5.4.
60 Tutorials
Figure 5.4 Plot of Oil FVF, Viscosity and Rs versus pressure for the output black oil property tables
Generating ECLIPSE Black Oil equilibration

keywords
This is similar to the generation of PVT tables. To generate equilibration tables, a
composition versus depth experiment is required.
2 Edit Experiments: Add | Composition with depth...
3 Click Next
4 In the General panel enter the reference properties for the sample:
• Enter 220 F as the Temperature.
• Enter 3580 psig as the Pressure.
• Enter 9200 ft. as the Depth
• Enter 0 F/ft. as the Temperature gradient.
5 In the Observations panel enter the depths 9000 ft. and 9400 ft.
6 Click Next
7 Click OK to allow PVTi to add extra points between the maximum and minimum
depths.
8 Click Close
9 Right-click on the new experiment (COMPG1) in the sample tree and select Export
keywords... from the drop-down menu.
10 Select the RSVD/RVVD (black oil) on the Equilibration Keyword radio button.
Tutorials 61
11 Click OK.
12 Enter the filename RSVD.PVO for the exported keyword.
Note In this case, only RSVD is generated. This is because the reservoir is all initially
in the liquid phase. If there were a gas-oil contact, both RSVD and RVVD would
have been generated. If the reservoir were all in the gas phase, only RVVD
would be generated.
13 PVTi: File | Exit.

Keywords have now been generated and can be incorporated into an ECLIPSE data-
set using ECLIPSE Office.
Importing the keywords into ECLIPSE Office

This section is not intended as a tutorial on using ECLIPSE Office. Refer to the
"ECLIPSE Office User Guide" for details on using the product.
1 Start ECLIPSE Office with a new case (call it PVTI_TUTORIAL.OFF) and import
the standard data set ECLIPSE100.DATA
2 Click ECLIPSE to let ECLIPSE Office know what type of data-set is being
imported.
3 Open the Data Manager.
4 Open the PVT Sections.
5 PVT Section: File | Import | Append..., and import the PVT table keyword file
(ECLIPSE100.PVO). Click OK to remove the warning message.
6 PVT Section: File | Import | Append..., and import the water keyword file
(PVTW.PVO). Click OK to remove the warning message
Note At this stage you may want to view the keywords or plot them. For details on
how to do this, refer the "ECLIPSE Office User Guide".
7 Close the PVT Section, saving the file with the new keywords.
8 Open the Initialisation Section
9 Initialisation Section: File | Import | Append..., and import the equilibration
keyword file (RSVD.PVO). Click OK to remove the warning message.
10 In the Equilibration Data Specification keyword, set the Rs/Pb v Depth Table to 1,
so that the imported RSVD keyword is used.
11 Close the Initialisation Section, saving the file with the new keywords.
12 Run the simulation.
62 Tutorials
Discussion
In this tutorial an existing fluid definition was imported into PVTi and exported as PVT
tables that were used in an ECLIPSE Black Oil reservoir simulation. The basic
requirements are that PVTi must simulate a Constant Volume Depletion or Differential
Liberation experiment and a Separator experiment to allow the generation of black oil
tables from the fluid definition.
Tutorials 63
64 Tutorials
Converting a black oil run to compositional
This tutorial provides an example conversion from ECLIPSE BlackOil to ECLIPSE
Compositional. This tutorial requires the use of ECLIPSE Office in combination with
PVTi.
Note Many conversion projects require conversion of SCHEDULE section keywords,

which is outside the scope of this tutorial. This tutorial covers conversion of
fluid properties and equilibration. Specifically, the PROPS and SOLUTION
sections.
Note Some familiarity with ECLIPSE Office is assumed for this tutorial.
The data for this project are provided with the standard installation of PVTi under the
directory:
• "Introduction" on page 65.
• "Exporting the fluid model" on page 66.
• "Converting equilibration keywords" on page 66.
• "Creating the ECLIPSE Compositional case" on page 67.
Introduction
In this tutorial, the black oil PVT tables (PVTO and PVDG) and the Equilibration table
(RSVD) are converted into a full compositional model and composition versus depth
table (ZMFVD). This allows the ECLIPSE data-set from the previous tutorials to be run
as a full compositional case.
Caution The --PVTi comments written out with the keywords are used by PVTi
to reconstruct the original fluid model. Without these there is not enough
information to convert blackoil projects to compositional models. It is
important that the lines prefixed by --PVTi in the ECLIPSE data-set are
not edited or moved around in the file.
1 Start ECLIPSE Office with the project created in "Exporting ECLIPSE Black Oil PVT
tables" on page 59 or create a new project and load in the data-set
ECLIPSE100_FULL.DATA.
2 Select the case.
3 Click on the PVTi launch button.
4 Click Run.
Tutorials 65
Tutorial 6: Converting a black oil run to compositional
Note The launch button has a default selection of launching PVTi with the PVT
section of the current case. PVTi reads this PVT section, creating a PVI file from
the --PVTi comments.
Exporting the fluid model

The imported PVT section contains the samples from the original PVI file plus any
experiments that were needed to generate the keyword. In this case the experiments
are a Differential Liberation experiment and a separator.
1 PVTi: File | Export Keywords | ECLIPSE Compositional Fluid Model...
2 Select the fluid {None}.
This means that PVTi does not write out a ZI keyword for the ECLIPSE
Compositional fluid model. This is the correct selection in this case as the
equilibration (RSVD) is used to create a composition versus depth table (ZMFVD).
3 Enter the reservoir temperature as 220o F.
Hint The reservoir temperature is the temperature in the Differential Liberation

experiment definition. You can right-click on the DL1 experiment and select
Edit... from the drop-down menu to view the definition of the Differential
Liberation experiment.
4 Click OK.
5 Export the fluid model to FLUID.PVO
6 PVTi: File | Exit (There is no need to save the PVI file as it can be created from the
ECLIPSE Office case).
Converting equilibration keywords

1 In ECLIPSE Office, click on the PVTi launch button.
2 Select Initialisation as the section to launch PVTi with.
3 Click Run.
Hint Again, PVTi searches for the --PVTi comments and uses them to construct a
PVI project file.
4 In PVTi, right-click on the composition versus depth experiment COMPG1.

5 Select Export keywords... from the drop-down menu.
6 In the COMPG1 export panel, select ZMFVD (Compositional) on the radio button.
7 Click OK.
8 Export the keyword to the file ZMFVD.PVO.
9 PVTi: File | Exit (there is no need to save the PVI file as it can be created from the
ECLIPSE Office case).
66 Tutorials
Creating the ECLIPSE Compositional case
1 ECLIPSE Office: Case | Add Case | Clone .
This creates an identical copy of the original case.
2 Name the case COMPOSITIONAL, and click OK.
3 Select the newly created case.
4 ECLIPSE Office: Module | Data Manager...
5 Select the Case Definition.
6 In the Case Definition module, select Compositional on the Simulator radio button.
7 Click OK to the warning about changing between black oil and compositional
cases.
8 In the PVT folder set the number of components to 11, and click OK.
9 In the Data Manager, select the PVT section.
10 PVT Section: File | Import | Append... and import the file FLUID.PVO.
11 PVT Section: File | Import | Append... and import the file ZMFVD.PVO.
12 PVT Section: Section | Keywords...
13 Delete the PVTO and PVDG keywords.
14 PVT Keywords: File | Close...
15 PVT Section: File | Close... and save the section with a new name.
16 In the Data Manager select the Initialisation section.
• Delete the RSVD keyword.
• In the EQUIL keyword set the Compositional init type to 1 (so that ZMFVD is
used for equilibration).
17 Initialisation Section: File | Close and save the section with a new name.
18 Run the simulation from the ECLIPSE Office Run Manager.
Discussion
In this tutorial, an ECLIPSE BlackOil simulation data-set was converted to ECLIPSE
Compositional using the integration of PVTi and ECLIPSE Office.
The insertion of the --PVTi comments into the keyword export from PVTi is a
powerful tool, not just for converting data-sets, but also for developing projects in
either black oil or compositional models.
Tutorials 67
68 Tutorials
Workflow Tutorial
Introduction
This tutorial illustrates a typical workflow for an oil or gas condensate. It involves
splitting the C7+ fraction into 2 pseudocomponents, special regression, normal
regressing, grouping components, and matching viscosity data.
We have an oil PVT case, with two fluid samples ZI and W2 (Well 2). There is a C7+
characterization with CO2 present. Well 2 has C7+ has a different MW and Specific
Gravity, but the C7+ has been characterized with the ZI fraction only at this point.
They are going to inject CO2 into this field, so there is a Swelling Test with CO2.
The files for this tutorial are provided in the default PVTi installation in the following
directory:
This tutorial contains the following sections:
• "Comparing the Default EOS Calculations to the Observations" on page 69.
• "Splitting the C7+ Component" on page 70.
• "Special regression to adjust the tail in the splitting calculation" on page 71.
• "Normal regression to fine tune 11 component match" on page 72.
• "Grouping Like Components to Reduce Nc" on page 72.
• "Regressing to match viscosities" on page 75.
Comparing the Default EOS Calculations to the

Observations
1 Start PVTi.
2 Open WORKFLOW.PVI.
3 Run the Simulations; to do this click GO.
4 Review the calculated and observed Bubble Point Pressures for fluid ZI and W2 in
the Output Display, that is the first and last experiments.
5 Close Output Display panel.
6 Plot the results one experiment at a time.
Right-click on CCE1 - Plot.
After reviewing the plots clear the plots by clicking the Remove Plots button.
7 Review all the experiments by observing the plots noting how well PVTi has done
in each case in matching the observations.
Tutorials 69
Tutorial 7: Workflow Tutorial
Splitting the C7+ Component
Creating a phase plot
Before we split we will create a Phase Plot (P versus T)
1 Select View: Samples| Phase plot | ZI; click OK.
2 Rescale the y-axis as follows:
• Double click on the y-axis.
• Select range.
• Click off Limit Range.
• Enter 0.0 in upper Visible Range area.
• Enter 3000 in lower Visible Range area.
• Click OK.
3 Click the “Superimpose“ button or select Options | Graph | Superimpose.
Splitting the C7+

4 Edit | Fluid Model | Split | Multi Feed....
We will split the C7+ into 2 Pseudo components.
5 Note the Mole Weight of the heaviest pseudo component.
6 Enter the following Specific Gravity and Molecular Weight of Samples Plus
Fraction for W2:
Molecular Weight 199

Specific Gravity 0.8338
7 Click OK.
The C7+ has been split.
8 Select Edit | Samples | Compositions.
Check the mole fractions of the 2 pseudo heavy components. The split creates
FRC2 with small mole fraction (0.0477). We would rather have a splitting that has
more of the mole fraction in the heaviest component so we will perform another
split.
9 Close this project and do not save the project.
10 Open WORKFLOW.PVI again.
11 Plot the Phase Diagram.
12 Select the Split panel and repeat step 6.
13 This time change the Mole Weight of Heaviest Pseudo Component to 300.
This gives us a larger mole fraction for FRC2.
14 Click OK.
15 Check the Samples.
The FRC2 mole fractions are 8.1% and 13.9%
70 Tutorials
16 Create a Phase Plot superimposed on the unsplit plot.
The phase diagram has not changed much at the reservoir temperature, which
indicates a good splitting.
17 Before we move on the regression save these results, using File | Save As, give the
file the name SPLIT.PVI.
Special regression to adjust the tail in the splitting

calculation
Before we regress we want to set the weights of the viscosity observations to 0, so they
are not included in the RMS. We will regress on the viscosity last, after we finish the
phase behavior match.
Hint Regressing on the viscosity coefficients after the phase behavior of the fluid
has been matched is always strongly recommended.
We will also increase the weights of the Bubble Point Pressure (Sat. pressure)
Observation as this is a very important criterion to match.
1 Right-click on Vapor Visc and enter 0 as the Set Weight.
2 Right-click on Liquid Visc and enter 0 as the Set Weight.
Note Setting the weight of an observation to zero turns off that observation.
3 Right-click on ZI | BUBBLE1 Z| Sat pressure and enter 40 as the Set Weight.

4 Right-click on W2 | BUBBLE2 | Sat pressure and enter 40 as the Set Weight.
5 Select Run| Regression.
6 Click Special and then Variables
7 Select all 3 SCT variables and press OK.
8 Select Regress | Run.
• Note the RMS in the Log area.
9 In the Regression panel click Simulation and observe the match of the 2 Bubble
Point pressure.
• Or alternately select Report | Regression.
10 Under Details check the matches of the observations, especially the 2 bubble point
pressures, both should match very well. Look at the Modifiers to see the amount of
change in the parameters.
Note You may also view the Sensitivities and the Hessian, Covariance and
Correlation matrices in the Report panel. These matrices are explained in the
"Reference section" on page 99.
Plotting the match

11 Switch off Superimpose, click Remove Plots.
Tutorials 71
12 Right-click on each experiment CCE1, etc. click plot, view the match and then click
Remove Plots before proceeding to the next experiment.
Note the DL results are much improved.
The general rule is if this special regression improves the match of the phase behavior
you should accept the regression. If it does not improve it or makes it worse reject the
regression. In this case the match is improved so we will accept this regression.
13 Click Accept and close the Regression panel.
14 Save the results, select File | Save As and save as ALPHA.PVI.
Normal regression to fine tune 11 component

match
You can do some normal regression to fine tune this 11-component match. The
Swelling and Separator Experiments still need to match. Or you can group now and
then do the normal regression to finish matching the phase behavior. This tutorial will
do the fine tuning using Normal Regression before grouping.
1 Set weights on bubble point pressures, etc. Right click on Observation | Set Weight
and:
• Make sure the 2 bubble point pressure observations weights are still 40.
• Set the Swelling Test | Saturation Pressure Weights to 10.
• Select CCE and set the Liquid Density Weights to 5.
• Set the Separator Exp Gas-Oil-Ratio weights to 5.
2 Select Run | Regression.
3 Select Normal under the Variables section.
4 Click Variables box.
5 Now refer to the section ""Normal Regression to fine tune match" on page 73 to fine
tune the match.
Note Do not attempt to get a perfect match since the grouping process described
below changes the match. The key here is to get the Swelling Test; Saturation
Pressures to match better before grouping.
6 Save this characterization with a new file name.
Grouping Like Components to Reduce Nc

We will group the 11 components into 6 components. Rules for grouping a fluid like
this are as follows:
• Keep Methane C1 as a pure component.
• Keep CO2 as a pure component - we will inject CO2.
• Keep the two heavy pseudocomponents that you created by splitting (and did
a special regression on) as separate components.
72 Tutorials
1 Before we group create a Phase diagram of component ZI and click the
Superimpose button.
2 Select Edit | Fluid Model |Group.
3 Type the following integers into the New Index column:
CO2 1
C1 2
C2 3
C3 3
IC4 3
NC4 3
IC5 4
NC5 4
C6 4
FRC1 5
FRC2 6
4 Click Update.
5 Check the component (Group) names. You change them if you do not like a
particular name.
6 Click Update again, and then click OK.
7 Click the Edit sample composition’s button to view the mole fractions of the new
group.
The usual rule is if one of the group mole fractions is significantly smaller than the
others group it into one of its neighbor groups.
8 Plot a Phase diagram on top of the un-grouped diagram.
If they are close to each other then the grouping is good. Your phase match should
still be close to the observed data.
9 Click GO and view the bubble point pressures
10 Plot the experiments to see how much the grouping has changed the calculated
results.
11 If the match looks reasonable save this characterization with a new file name.
12 If the match has changed substantially, close this project without saving and read
in the .PVI file from the save at the end of the previous section.
13 Re-group with a different selection of groups or number of pseudocomponents.
Normal Regression to fine tune match

The steps in the normal regression process are:
• Choose high weights on experiments or observations to improve the of key data
• Pick various combinations of parameters to regress on, try it, look at the results,
reject the regression and try a different combination
Tutorials 73
• Volume shift parameters may be dependent on Tc and Pc - so you cannot regress
on them separately. They may be independent in which case you can regress on
them. If you are having trouble matching liquid densities try making the volume
shifts independent.
A review of the match of the observations shows that the Saturation Pressure for the
Swelling Experiment with CO2 contains the largest difference between the calculated
and observed. We will increase the weight for these results.
There are two ways to change the weight for experiments and observations:
• To change the global weight for the experiment:
• Right -click on the experiment and click on set weight -
• Type in the new weight.
• To change the weight on individual observations :
• select Edit | Observations:
• click on the experiment type,
• click experiment list,
• then observation type,
• then click on the large G in the upper left corner of the panel.
Individual weights appear on the right most columns.
• Insert weight values and click Apply.
Suggested weight values:
• Swelling, saturation pressure = 10.
• CCE, liquid densities = 5.
• SEPS, Gas-Oil Ratio = 5.
• Bubble point pressure = 40 (previously set).
Regression variable trial and error process

For the remainder of the phase behavior regression, it is a trial and error procedure.
Below are several suggested combinations of parameters to regress on. Remember a
vertical column of numbers in the Regression panel:
1
1
1
1
creates one regression variable for all four components.
To create 4 regression variables one should enter the following in the column:
1
2
3
4
74 Tutorials
Below are 3 combinations of suggested variables (with volume shift parameters
"Independent"). A variable should be defined for the contents of each bullet point in
the 3 combinations:
• Omega A - C9+, C23+
• Omega B - C9+, C23+
• AcenFac - C9+, C23+
• Shift - all components
• BIC CO2 and C1, C2+, C5+, C9+, C23+
• BIC C1 and C5+, C9+, C23+
or:
• Omega A - C2+, C5+, C9+, C23+
• Omega B - C2+, C5+, C9+, C23+
• AcenFac - C2+, C5+, C9+, C23+
• Shift - all components
• BIC CO2 and C1, C2+, C5+, C9+, C23+
• BIC C1 and C5+, C9+, C23+
or:
• Tc - C9+, C23+
• Pc - C9+, C23+
• AcenFac - C9+, C23+
• Sshift - all components
• BIC CO2 and C1, C2+, C5+, C9+, C23+
• BIC C1 and C5+, C9+, C23+
1 Try as many combinations as required to match the data to within the degree of
accuracy you think is needed. Remember the accuracy of most PVT observations
is 5% to 10% or about 20 Psi.
2 Once you have finished matching the phase behavior data, accept the results of the
regression and save the file.
Regressing to match viscosities

Now that we have a match of the phase behavior, we next need to match the viscosity
observations. First we need to remove the experiments that do not have viscosity data
from the regression process. Then in the CCE we need to set the viscosity observations
weights to 1 and the other observations to 0.
1 To remove an experiment from regression right click on the experiment and select
Don't use in regression. Do this for all experiments except CCE.
2 In the CCE experiment for all the observations (except for the Vapor visc. and
Liquid visc.) right click on the observation and select set weight.
• Change the weight to 0.
Tutorials 75
3 For the Vapor visc. and Liquid visc. change the weight (using the same sequence)
from 0 to 1.
4 Select Run | Regression.
5 In the Variable section click Normal then the Variables box.
First iteration
6 In the Select EOS parameters for regression panel enter 1 in all the boxes under
the heading VcritV.
7 Click OK.
8 In the Regression panel click Run.
9 Plot the Liquid and Vapor viscosity and compare the new match with the
observations.
Second iteration
10 In the Regression panel press Reject to return the characterization back to the pre-
regression values.
11 In the Select EOS parameters for regression panel enter 1 in the first box, 2 in the
second box, 3 through 11 in the remainder of boxes under the heading VcritV.
12 Click OK.
13 In the Regression panel click Run.
14 Plot the Liquid and Vapor viscosity and compare the new match with the
observations.
Note this new match is better than the first match.
15 Press Accept to accept these results.
16 Save this characterization with a new file name.
The phase matching process is now complete. You are ready to export the PVT
properties or characterization for ECLIPSE simulations.
Discussion
This tutorial illustrated a typical workflow for an oil or gas condensate. It involved
splitting the C7+ fraction into 2 pseudocomponents, special regression, normal
regressing, grouping components, and matching viscosity data.
76 Tutorials
Multiphase Flash
Introduction
The multiphase flash experiment tends to find more than two phases in systems with
Asphaltene/Waxes and/or with CO2 rich fluids at low temperatures. This tutorial
demonstrates multiphase flashes with both systems. The files for this tutorial are
provided in the default PVTi installation in the following directory:
The tutorial is split into the following sections:
• "Asphaltene and wax system" on page 77.
• "CO2 Rich Fluids" on page 78.
• "Summary" on page 80.
Asphaltene and wax system

2 Load MULTIPHASE-START1.PVI into PVTi.
3 View this oil composition by selecting Edit | Fundamentals and view the 10
component oil.When finished click OK.
4 To add an experiment, click Edit | Experiments | Add | Single Point | Multiphase
Flash.
• Click Next> and then enter 50 F as the Temperature,
• Click Observations, fill in 1000 Psia.
• Click Next> and then Close.
5 To run the experiment (again) click GO.
The Output Display shows the results of the flash.You will see two phases, Liquid
and Vapor, and their properties and compositions.
Now we are going to split the C7+ fraction into it Paraffin, Naphthalene, and Aromatic
components, then redo the multiphase flash.
6 Select Edit | Fluid Model |Split | PNA Distribution.
The C7+ fraction is now split.
7 To view the new characterization, select Edit | Fluid Model Components.
You will see that there are three new user-defined components which have
replaced the C7+ component.
8 Click on the Complete folder to view the critical properties of these components.
9 Click on OK to close the panel.
We are going to run the MFLASH experiment again and view the phases present.
10 Click on GO .
Tutorials 77
Tutorial 8: Multiphase Flash
The Output Display shows the following 4 phases: Liquid, Wax, Asphaltene Liquid,
and Vapor.
11 Note that the compositions of the Wax is 100% PC7+ and the Asphaltene Liquid is
90.23% AC7+.
12 Close the project, do not save the changes.
CO2 Rich Fluids

Certain fluids with a high CO2 content at low reservoir temperatures partition into two
liquid phases or two liquid phases in equilibrium with a vapor phase. This tutorial
demonstrates such a system.
SPE 71485, (see [Ref. 1]) gives fluid characterizations that exhibit multiphase behavior.
This paper describes reservoir oil with 12 components. It has heavy components of C7-
9, C10-13, C14-19, C20-35, and C36+. It also describes an injection gas called (MI,
Miscible Injection) that is a combination of CO2 and NGL. In the paper they use the
Peng-Robinson Equation-of-State to calculate the phase behavior. They combine the
reservoir oil and the MI gas in various mixtures at 86 °F and present a diagram of the
phases present, which is shown in Figure 5.5.
Figure 5.5 Phase Diagram for Schrader Bluff Fluids
We will attempt to verify the phases present with a 0.8 fraction of MI with PVTi
multiphase flash.
1 Load MULTIPHASE-CO2.PVI into PVTi.
This contains the fluids and characterization from the SPE paper,[Ref. 1].
2 View the compositions of the fluid sample by selecting Edit |Samples -
Compositions.
78 Tutorials
3 Click OK to close the panel.
4 To create a mixture of 80% MI and 20% reservoir oil, select Edit | Sample | Mix.
5 In the Mix panel enter the following:
• Mixing Type By Mole Fraction of Sample 2

• Fluid Sample 1 Z1
• Fluid Sample 2 MI
• New Sample Name .8MI
• Temperature 86 F
• Mole Fraction 80 percent
• Click OK.
6 To view the new sample, click Edit | Samples | Compositions.
Note the new sample has 65.209% CO2.
Now we will create multiphase flash experiments at a series of pressures (Temperature
= 860 F) starting in the Liquid-Liquid region (1100 Psia) then through the Liquid-
Liquid-Vapor region and ending up in the Liquid-Vapor region (600 Psia).
7 Select Edit | Experiments | Add | Single Point | Multiphase Flash.
8 Enter the following:
• Fluid Sample .8 MI
• Temperature 86 F
9 Select Observations and enter 1100 (psia) as the Pressure = 1100 (psia).
10 Click Next> and Close.
We now have MFLASH1 defined.
11 To create additional MFLASH experiments at a series of lower pressure, right click
on MFLASH1 and select Clone.
12 Repeat for MFLASH2 through to MFLASH5.
We now have 5 MFLASH experiments defined.
13 To change the flash pressures right click on the MFLASH experiment and select
Edit | Observations | Pressure.
14 Enter the following pressure values:
MFLASH2 1050 (psia)

MFLASH3 1000 (psia)
MFLASH4 900 (psia)
MFLASH5 600 (psia)
• To switch between MFLASH experiments press Next> and then Close.
15 To run all experiments and view results, click GO.
16 Observe the results in the Output Display.
Note as the pressure decreases the flashes proceeds from the L-L region to the L-L-
V region to the L-V regions, just as Figure 5.5 illustrates.
Tutorials 79
Note The MFLASH5 experiment can sometimes label both the phases as liquid.
However, one of them is clearly a vapor as the density value is 5.64323 lb./ft.3.
Note If a standard two-phase flash is performed at the same temperature and

pressure as with the multiphase flash, then one obtains liquid and vapor
phases with the same density values as produced with the MFLASH5
experiment.
Summary
This tutorial demonstrated the multiphase flash experiment. It tends to find more than
two phases in systems with Asphaltene/Waxes and/or with CO2 rich fluids at low
temperatures. This tutorial demonstrates multiphase flashes with both systems.
References
Guler B. et al, "Three- and Four-Phase Flow Compositional Simulations of CO2/NGL EOR" [Ref. 1]
SPE 71485
80 Tutorials
Exporting an ECLIPSE Thermal model
Introduction
This tutorial demonstrates using the new ECLIPSE Thermal export facility where a file
can be exported containing a fluid model suitable for use in ECLIPSE Thermal. For
technical information on the ELCLIPSE Thermal export facility see "ECLIPSE Thermal
Export Module" on page 423 and for more general workflow guidelines see
"Compositional Data for ECLIPSE Thermal" on page 387. The files for this tutorial are
provided in the default PVTi installation in the following directory:
The tutorial is split into the following sections:
• "Verifying the Validity of the Fluid Model" on page 81.
• "Fitting the Component K-values" on page 82.
• "Viewing the K-value Fits" on page 83.
• "Exporting the Model" on page 84.]
Verifying the Validity of the Fluid Model

2 Open THERMAL.PVI.
3 Open the Fundamentals panel by selecting Edit | Fundamentals....
From the tree view on the left side of PVTi you can see that there is a single sample
in this project called ZI. The Fundamentals panel shows the composition of this 3-
component fluid as being C1, C5 and C20+.
4 On the tree view there are two experiments defined, a Differential Liberation (DL1)
and a bubble point experiment (BUBBLE1). Right-click on the DL1 experiment and
select Plot.
Three observations should have been plotted oil density, oil relative volume and
oil viscosity. You can see that the Equation of State (EoS) fluid model shows good
agreement with all 3 observations.
5 Right-click on the bubble point experiment and select Report.
You can see that the EoS model also gives good agreement with the observed
bubble point pressure of 1784.1749psia.
Since we have a good match for our EoS based fluid model we can now export the
model for use in an ECLIPSE simulation. We have relatively few components (<4)
so this fluid would be suitable for use in an ECLIPSE Thermal simulation.
Note Simulations using ECLIPSE Thermal tend to use fluids consisting of two or
three components.
Tutorials 81
Tutorial 9: Exporting an ECLIPSE Thermal model
Fitting the Component K-values
1 Right-click on the sample ZI and select the Export ECLIPSE Thermal model...
2 On the panel that opens, enter the following:
• 1500 psia as the Maximum Pressure,
• 400 F for the Maximum Temperature,
• 1000 psia for the Minimum Pressure,
• and 200 F in the Minimum Temperature box.
Note The default values here are Pmax=1000psia, Pmin=50psia, Tmax=400F and
Tmin=50F and are considered reasonable max/min parameters within a
reservoir. However, every reservoir is different and any knowledge of these
parameters for your particular reservoir should be entered.
3 Enter 40 in the Enter Number of Flashes to be Performed box.

To model component K-values we can either export the KVWI keyword, which models
them using Wilson’s formula, or the KVCR keyword, which uses Crookston’s formula.
Crookston’s formula is in general much more accurate and we will use this. See "K-
Values" on page 423 for a more detailed description regarding the modeling of K-
values.
4 Tick the box Export Crookston Coefficient? to tell PVTi that you wish to export the
KVCR keyword.
5 Since we are exporting the KVCR keyword we need to determine the values of the
coefficients of Crookston’s equation to export. Click Fit Crookston Coefficients on
the panel to open the Fit Crookston Coefficients panel.
This panel shows Crookston’s equation where p is the pressure, T is temperature
and the coefficients A-E are what we wish to determine values for.
Note The Fit Crookston Coefficients panel enables you to calculate the optimum
values of the coefficients A-E in Crookston’s formula, so that the best fit is
found to PVTi’s EoS predicted K-values for each component over the
temperature and pressure range defined by the user.
6 A and D should already be active. Click on B to make coefficient B active also.

7 Select the Plot option in the Plot P, T Values Used in Fitting Crookston Coefficient?
box.
In order to find values for the chosen coefficients PVTi throws in 40 points at random
coordinates in the region you just defined in pressure-temperature space. The Plot
option plots these points on the screen for you after the fit has been performed. Ideally,
we would like to them superimposed on a phase plot.
8 Select PVTi: View | Samples | Phase Plot... and press OK to perform the phase plot.
9 Select PVTi: Options | Graph | Superimpose.
Hint You can also access the Superimpose option using the toolbar.
82 Tutorials
10 Now click Apply on the Fit Crookston Coefficients panel to start the run.
Once the run has finished a results panel appears. In the Coefficients folder the best
fit values of A, B and D are reported for each component.
11 Click on the Statistics panel.
The mean error and standard deviation (in %) are reported for each fit. The C1 and
C5 components have been fitted very well (rms<1.5%) and the C20+ fraction has
been fitted reasonably well with an rms of somewhere between 7-9% (depending
on the random number generator on your machine).
Can we do better though?
Caution Make sure you turn off the Superimpose option before moving to the next
section.
Viewing the K-value Fits

In the last section "Fitting the Component K-values" on page 82 we saw how to use the
module to calculate the optimum values of a chosen set of coefficients in Crookston’s
equation in terms of fitting to PVTi’s Equation of State based K-values. We saw, for the
fluid ZI in the THERMAL.PVI file, that the C20+ fraction had a reasonable fit when
using just A, B and D. In this section we will see how to interactively view the fits in
order to better understand why PVTi’s EoS K-values for this C20+ fraction has not been
represented as well as the other components.
1 Click the View Fit button on the Fit Crookston Coefficients panel.
The Plot K-values vs Temperature or Pressure panel opens.
Hint Plots can either be performed of K-values versus pressure (at constant
temperature) or K-values versus temperature (at constant pressure).
First we will look at the K-value versus temperature fits, which are dictated by the
D and E coefficients (just D in our case).
2 Enter 1250psia in the Enter Constant Pressure box and 400F and 200F as the
Maximum and Minimum Temperatures respectively. Now click Apply.
The PVTi EoS-based K-values are shown by the points and the K-values calculated
using Crookston’s equation are shown by the curves.
3 Experiment by changing the value of the constant pressure in the range
1000<P<1500 (our chosen pressure range) to see how well Crookston’s formula
models the temperature dependence of the K-values at a given pressure.
In general, for the C20+ component, the D coefficient, models the observations
pretty well over the whole pressure-temperature range, although using the E
coefficient as well may well help slightly
4 Click the Plot at Constant Temperature box and enter 300F.
5 Now enter the appropriate pressure range, which is 1500 psia and 1000 psia for
the maximum and minimum values. Click Apply.
Tutorials 83
6 Again, experiment by changing the value of the constant temperature in the range
200<T<400 to see how well Crookston’s formula models the pressure dependence
of K-values.
In particular it can be seen in the range 300<T<400 (for the C20+ component) that
Crookston’s formula struggles to model the pressure dependence at pressures at
P<1100psia when using just the A and B coefficients. Using the C coefficient may
well improve things significantly.
You can see in particular, the K-value versus pressure curve for the C20+
component struggles to fit the observations.The problem is that the term A+B/P
starts to run into problems for pressure values <1100psia (due to the strongly
negative value of B) and therefore we require the C coefficient to get a good fit.
7 Close the Crookston Report panel and the Plot K-values vs Temperature or
Pressure? panel.
8 Switch on the C and E coefficients by ticking the appropriate boxes on the Fit
Crookston Coefficients panel. Now click Apply.
This run will take slightly longer, as the introduction of the C and E terms vastly
increases the amount of parameter space that PVTi must search.
9 Once the run has finished repeat steps 1-3.
This time, due to the introduction of the C coefficient in the fit, we have done a
better job in fitting the pressure dependence of Crookston’s equation. The
introduction of the E coefficient has also slightly improved the modeling of the
C20+ K-value versus temperature.
10 Click on the Statistics folder.
Overall, the introduction of the C and E coefficients has decreased the rms fit for
the C20+ component from ~8% to ~5%.
Exporting the Model

Now we are satisfied with our K-value fits we can export our ECLIPSE Thermal PVT
model.
Note If we had decided to export the KVWI keyword (that is use Wilson’s formula to
model K-values) then we would not have needed to fit the coefficients of
Crookston’s equation and could have exported straight away from the Export
for ECLIPSE Thermal panel. Although this may be quicker, Crookston’s
formula models K-values much better and spending this extra time is
worthwhile.
1 Click the OK button on the Export for ECLIPSE Thermal panel.

2 Choose a name for the file to be exported.
By default PVTi names it rootname.PVOT, so in this case THERMAL.PVOT if you
do not choose otherwise.
3 Click Save.
The ECLIPSE Thermal fluid model is written to the specified file. This file can now
be used to model PVT behavior as part of an ECLIPSE Thermal simulation.
84 Tutorials
Data analysis and quality control
This tutorial provides a typical workflow for PVTi in its role as a data quality
assessment tool. Experimental results from analysis of a hydrocarbon gas is used to
analyze the data quality and to modify spurious data. The data for this project are
provided with the standard installation of PVTi under the directory:
• "Material balance checking" on page 86
• "The Hoffman-Crump-Hocott test for separator data" on page 90
• "Recovery calculations" on page 91.
Introduction
In addition to allowing an equation of state to be fitted to laboratory results and
facilitating the generation of ECLIPSE BlackOil/ Compositional PVT data, PVTi also
provides material balance checks to assess data quality. For information on the
calculations involved in material balance checking see "Compositional material
balance" on page 320.
Note Problems with the observations in a PVT report equate to problems with the
fitted fluid model. It is therefore recommended that material balance checks
are carried out on all PVT data.
In this tutorial an existing project file (GAS.PVI) is read into PVTi and the data are
checked and modified for material balance errors.
2 PVTi: File | Open...
3 Open GAS.PVI
Hint The Data Tree should show that there are two experiments, CVD1, defined
with 10 different types of observations, SEPS1 with observations of fluid mole
fractions (liquid and vapor) and CCE1 with observation of Vapor Z-Factor.
4 Click and drag the ZI node from the Data Tree and drop it into the Main Plot
Window.
The phase envelope should look like Figure 5.6.
Tutorials 85
Tutorial 10: Data analysis and quality control
Figure 5.6 The phase envelope plot.
Note This fluid system has no well-defined critical point.
Material balance checking

1 Right-click on CVD1 in the sample tree and select Material Balance...
This opens the Material Balance panel for this experiment.
2 Click Report to create a material balance report.
The experiment is performed and the Output Display window opens showing
messages that indicate the quality of the data. (Figure 5.7)
86 Tutorials
Figure 5.7 The main display shows messages indicating the quality of the data
Warning - Sg of final stage liquid plus fraction is not defined

Warning - Mw of final stage liquid plus fraction is not defined
Warning - Viscosities of gas of at least one stage of CVD is not defined
Warning - Mw of vapor plus fraction of at least one stage is not defined
- setting constant Mw(CN+) = of Mw(CN+) at Psat
Warning - Sg of vapor plus fraction of at least one stage is not defined
Warning - Composition of final stage liquid does not sum to 100%
Calculated liquid mole% of N2 at P= 6300.00000 is negative
Calculated liquid mole% of CO2 at P= 6300.00000 is negative
Calculated liquid mole% of IC4 at P= 6300.00000 is negative
Calculated liquid mole% of NC4 at P= 6300.00000 is negative
Calculated liquid mole% of IC5 at P= 6300.00000 is negative
Calculated liquid mole% of NC5 at P= 6300.00000 is negative
Calculated liquid mole% of C6 at P= 6300.00000 is negative
The messages show that some mole fractions were calculated as negative, so there
are clearly problems with the data. PVTi supplies various options for plotting the
data to ascertain the source of the errors. The first type of data check to perform is
to view the pressure variation of the gas compositions.
4 Click Plot in the Material Balance panel.
• Select Vapor Compositions v Pressure in the Select Plot Type panel and click
on Plot.
• Click Close
5 PVTi: View | Rubberband Zoom In

6 Click and drag the mouse to define the zoom area to approximately cover the
region 2800 to 6500 psia and 0.1 to 2 vapor composition. After zooming in, the plot
window should look similar to Figure 5.8.
Tutorials 87
Figure 5.8 The main plot window after zooming in
Many of the components have non-monotonically varying gas compositions. In

general, there are several fluids or analyses available, and bad data can be discarded.
However, if no other data is available PVTi offers tools to make modifications to the
bad data.
Modifying CVD data

7 Click Modify in the Material Balance panel.
8 Select fraction modifiers.
9 Enter the following values in for 6996 psia:
Component N2 CO2 ... IC4 NC4 IC5 NC5 C6

Percentage 10 10 ... 2 10 5 5 20
Note The other components are modified in proportion to their existing mole
fractions. Placing a letter in the thin column to the left of each column of
modifiers allows the proportion of that component to be fixed and thus not
modified in proportion to its existing mole fraction.
10 Click on OK in the Set correction factors for CVD compositions panel.

11 Click Report in the Material Balance panel to create a new material balance report.
12 Click on Yes to modify the compositions on CVD vapor and Liquid Composition
Modif... panel.
88 Tutorials
13 Click on No on the CVD vapor and Liquid Composition to ... panel so that the
compositions are saved, but retain their original values until after the modified
results have been examined.
Now none of the liquid mole percentages are negative. So this change to the data
can be accepted.
14 Click Report in the Material Balance Panel
15 Click on Yes to modify the compositions.
16 This time, click on Yes so that the compositions are modified.
Plotting K values versus Pressure

17 Click Plot in the Material Balance panel.
18 Select K-values:(1) log (k) v Pressure plot in the Select Plot Type panel and click
on Plot.
The plot window should now look like Figure 5.9.
Figure 5.9 The plot of k values versus pressure.
The K-values should plot monotonically in that N2 should be the largest, followed by
C1, etc. This is clearly not the case, so although there are now no calculated negative
compositions, the modified fluid definition is not fully consistent.
The Hoffman-Crump plot

19 Select the K-values:(2) Hoffman-Crump Plot in the Select Plot Type panel and click
on Plot.
20 Click Close in the plot panel.
21 Click Close in the Material Balance Panel.
Tutorials 89
The plot window should look like Figure 5.10.
Figure 5.10 The Hoffman-Crump plot
In this plot, one line is generated for each pressure stage. The Hoffman F
coefficients correspond to C1, C2 etc. and the lowest to C11, C12+. In general, these
lines should be monotonic with pressure, with the highest pressure at the top. This
plot shows most of the error to be in the first stage.
The Hoffman-Crump-Hocott test for separator data

Applying the Hoffman-Crump-Hocott test to separator gas and oil samples indicates
whether or not the streams are genuine equilibrium fluids.
1 Right-click on SEPS1 in the project tree-view and select Recombination... on the
pop-up menu.
2 Click Report to create a recombination report.
4 Click Plot.
The two lines on the Hoffman-Crump-Hocott plot (Figure 5.11) show the actual
data and the Standing estimates of K-values. They are used as a consistency check
and, in this case, give further evidence that the initial feed stream composition is
in error.
90 Tutorials
Figure 5.11 Hoffman-Crump-Hocott plot.
Recovery calculations
PVTi can allow recovery calculations to be performed if a valid Constant Composition
Expansion, Constant Volume Depletion and Separator test exist.
1 Right-click on CCE1 in the project tree-view and select Recovery... on the pop-up
menu.
2 Click on Report to perform the recovery calculation.
Hint This assumes that there is no direct production of reservoir liquid. If you want
to include direct production of reservoir liquid, you need a relative
permeability table, which you can enter be clicking on Rel. Perm.
Note PVTi runs the material balance check on the Constant Volume Depletion
experiment selected, and performs recombination on the Separator selected,
before performing the recovery calculation.
Discussion
This tutorial has illustrated how fluids may be examined for consistency and, if
necessary, modified, within a PVTi project.
Tutorials 91
92 Tutorials
Removing contamination from samples
Introduction
Oil-based muds are in widespread use and often contaminate PVT samples taken at the
well-site.
This tutorial involves the cleaning of a sample that is contaminated by an oil-based
mud.
The PVI file CLEAN.PVI is used for this tutorial and is provided in the default PVTi
installation in the following directory:
• "Removing oil-based mud contamination by skimming" on page 93
• "Removing oil-based mud contamination by subtraction" on page 94
Removing oil-based mud contamination by

skimming
2 Open CLEAN.PVI.
3 Right-click on the sample ZI and select Fingerprint plot from the drop-down menu.
In naturally occurring hydrocarbons there is expected to be semi-log straight-line
behaviour for components C8+ (around a mole weight of 100). From the
fingerprint plot, there is clearly not straight-line behavior for this fluid.
The contaminating mud, like many oil-based muds, has a composition containing
components C10-C23. In the skimming method, it is assumed that the composition
is not known.
4 Right-click on the sample ZI and select Clean... from the drop-down menu.
• Enter CLEAN as the sample name for the cleaned sample.
• Enter CONTAM as the sample name for the contaminant.
• Click OK.
The sample has now been cleaned.
5 PVTi: Options | Graph | Superimpose - and ensure that the Superimpose option is
on.
6 Right-click on the sample CLEAN and select Fingerprint plot from the drop-down
menu.
Tutorials 93
Tutorial 11: Removing contamination from samples
7 Right-click on the sample CONTAM and select Fingerprint plot from the drop-down
menu.
The plot should now look like Figure 5.12.
Figure 5.12 The original sample, the cleaned sample and the estimated contaminant.

subtraction
When the composition of the contaminant is known, the subtraction method can give
better results than the simple skimming method.
1 Right-click on the sample MUD and select Fingerprint plot from the drop-down
menu.
The true composition of the contaminant contains components lighter than C8 and
also up to the plus-fraction (C25+). The skimming method could not remove this
contaminant completely, but the subtraction method can.
94 Tutorials
2 Right-click on the sample ZI and select Clean....
• Enter CLEAN2 as the sample name for the cleaned sample.
• Select Subtraction as the method.
• Select MUD in the Contaminant drop-down.
• Click OK.
3 Right-click on the sample CLEAN2 and select Fingerprint plot from the drop-down
menu.
Discussion
This tutorial showed how a fluid can be cleaned of oil-based contaminants such as
drilling muds. For information on how the skimming and subtraction methods work
see "Removing contamination from samples" on page 93.
In general different PVT samples contain different levels of contaminant. It is usually
best to fit the PVT reports from a number of (contaminated) samples. Once a consistent
fluid model has been developed, the samples can be cleaned using either of the
methods outlined in this tutorial. The cleaned samples can then be used in reservoir
simulations.
Tutorials 95
96 Tutorials
Converting old projects to the current version
This tutorial demonstrates conversion of an old project to the current version of PVTi.
This is especially important for projects (PVI files) created with versions before 99B, as
the default Field units for Gas Formation Volume Factor were changed for that release.
The files for this tutorial are provided in the default PVTi installation in the following
directory:
• "Preparing the PVI file for conversion" on page 97
• "Converting the file" on page 98
Introduction
The VERSION keyword, introduced in 2000A, allows a systematic method for updating
old PVI files to be compatible with the latest version of PVTi. This tutorial describes
how the keyword can be used to convert an old PVI file into the current version.
Caution Files in FIELD units containing Differential Liberation (DL) experiments

that have Gas formation volume factor (GFVF) observations must be
updated to the current version.
Preparing the PVI file for conversion

1 Start PVTi with a new project.
2 Click Cancel in the Fundamentals panel.
3 PVTi: Utilities | Text Editor
4 Select OLD.PVI as the file to be viewed
5 In OLD.PVI enter the VERSION keyword in the RUNSPEC section with a value of
98B.
Hint If you are unsure of the form of the VERSION keyword see "VERSION
Version of PVTi" on page 291.
6 File | Save As...

7 Save the file with the name CONVERT.PVI
8 File | Close
Tutorials 97
Tutorial 12: Converting old projects to the current version
Note You could now use the file CONVERT.PVI in a normal session and PVTi
interprets it according to the version specified by the VERSION keyword.
In this tutorial we go one step further and convert the PVI file to the current version.
Converting the file

1 PVTi: File | Open
2 Select CONVERT.PVI as the file to open.
3 PVTi: File | Save As...
4 Save the file with the name NEW.PVI
Hint You can compare the file NEW.PVI to OLD.PVI to see the differences (the DL
observation GFVF is converted from rb/stb to rb/Mscf and the heat capacity
keywords are added in the SYSTEM section).
Discussion
In this tutorial an old PVI file was converted to the latest version. This is important for
files using FIELD units, containing Differential Liberation (DL) experiments that have
Gas formation volume factor (GFVF) measurements as the units, for this type of
observation was changed in 99B from rb/stb to rb/Mscf to make PVTi’s units systems
consistent with those of the ECLIPSE simulators.
98 Tutorials
Tutorial 12: Converting old projects to the current version
Reference section
Chapter 6
General information
• "Main PVTi window" on page 100
• "File" on page 102.
• "View" on page 105
• "The fluid model" on page 109
• "COMB - Compositional Material Balance" on page 123
• "Simulation using PVTi" on page 128
• "Regression in PVTi" on page 137
• "Exporting keywords" on page 144.
• "Utilities" on page 155.
• "Batch system and keywords" on page 163.
• "Error handling" on page 174.
Reference section 99
General information
Main PVTi window
General information
PVT analysis involves fitting an Equation of State to experimental data and then using
the Equation of State to produce PVT tables for use in reservoir simulations. PVTi
contains facilities to allow you to import experimental data, fit the data to an Equation
of State, and finally produce the PVT tables for reservoir simulation studies.
The menu bar of the main PVTi window has the following options:
• "File" on page 102.
• "Edit" on page 103
• "View" on page 105
• "Run" on page 107
• "Utilities" on page 107
• "Graph" on page 108
• "Window" on page 161.
• "Help" on page 161.
100 Reference Section

Main PVTi window
The PVTi main module
The main module is shown in Figure 6.1.
Figure 6.1 The main PVTi window
Data Tree Log Window

Equation of State
Main Plot Window and Sub-plots
The main window contains all the tools necessary for Equation of State model fitting.
Basic features
The Data Tree provides a view of the current project’s contents. Each fluid sample is
identified with its experiments as sub-nodes in the tree. Likewise, each experiment has
its observations as sub-nodes. The Log Window is updated with pertinent information
relating to actions taken in PVTi. The Equation of State, upon which the current fluid
model is based, is indicated in the status bar. The Main Plot Window and the Sub-plots
provide an area for viewing project information graphically.
Reference Section 101

Main PVTi window
File
The File menu allows you to open, close and save PVTi project files (PVI files) and
import sections from PVI files, and provides access to keyword export modules. Graph
printing and plotting facilities are also available from this menu.
• To open this menu, select File from the main PVTi window.
The File menu consists of the following options:
• New...
This creates a new PVTi project.
• Open...
Opens a PVTi project (PVI) file. The complete file is read in and the most recent
fluid model, experiment descriptions, observations, etc., are restored. For more
information on the files PVTi creates see "Files created by PVTi" on page 103.
• Close...
Closes the current project. If the project is not empty you are asked if you want to
save before closing.
• Save...
Saves the current PVI file, overwriting the previously saved project.
• Save As...
Saves the current project to a new PVI file.
• Save (concise)...
Saves a concise version of the current project containing the latest version of the
fluid model plus any experiments and observations used in simulation and
regression. No other information is saved, therefore information regarding
regression variables or split/group sections will not be recorded by the Save
(concise) option.
• Export Keywords
This provides access to the Keyword Export modules. Currently PVTi supports
export for the ECLIPSE simulators and VFPi. See "Exporting keywords" on page
144.
• Import
This option allows a section from a previous PVI project or ECLIPSE data-set to be
imported or a “concise” PVI project to be merged with the current project. The
sections that can be imported here are SYSTEM,GROUP, SPLIT,SIMULATE and
REGRESS. See "Reading the SYSTEM section from a PVI or DATA File" on page
109.
• View PVI Section
This opens a particular section from a PVI file and displays the keywords in a text
editor. The sections that can be viewed this way are SYSTEM, GROUP, SPLIT,
SIMULATE, and REGRESS. See "Displaying the SYSTEM section from a PVI file" on
page 109.
• Exit
This exits PVTi. If there is an active project, you are asked whether you would like
to save the project before exiting.

File
Files created by PVTi
All files in PVTi use the project name as their base name. PVTi creates the following
files:
• PVI file, for example ALL.PVI
Input data file, although can be written by PVTi to save system specification or
session.
• PVP file, for example ALL.PVP
Main printed output file. In interactive mode a prompt to write results to this file
follows most operations.
• PVO file, for example ALL.PVO
Output file. Used for the output of ECLIPSE Black Oil, GI option (pseudo-
compositional) or ECLIPSE 300 (equation of state) properties.
• VEC file, for example ALL.VEC
Vectors file. Contains vectors of plots performed in a PVTi section in a form
suitable for inclusion into the GRAF program.
• DBG file, for example ALL.DBG
Debug file. This is only present if debug has been written.
• MES file, for example ALL.MES
Message file. A temporary file used throughout the program run to display results.
This file is deleted when you quit the program.
• LOG files, for example ALL.LOG
Program Log File. This file exists in the startup directory of the program and
contains a summary of keywords read in, tasks performed, etc.
• NEW files, for example ALL.NEW
New data file. This is a temporary file that holds the details of the new .PVI file.
It is left in the working directory if the program does not shut down cleanly.
Hint The .NEW file contains all the changes made during the last session. If you
change the file extension to .PVI you can use it to recover the session.
• REG files, for example ALL.REG

Regress Module file. This temporary file holds details of the quantities plotted in
the Regress module. It is left in the working directory if the program does not shut
down cleanly.
Only one project at a time can be in use with a single run of PVTi. To open another
project, close the current project, either by selecting the File | Open option (the program
prompts you save the session to a new .PVI file) or by using File | Close.
Note Although only one project may be in use by the program, different sections of
different .PVI files may be read in.
Edit
The Edit menu allows entry and editing of the fluid model, samples, experiments,
observations and regression variables.

Edit
• To open this menu, select Edit from the main PVTi window.
The Edit menu consists of the following options:
• Fundamentals...
This opens the Fundamentals panel, See "Fluid Properties Estimation" on page 35.
• Fluid Model
This opens the sub-menu of fluid model editing options.
• Equation of State...
This opens the Equation of State selection panel. See "Equation of State" on
page 111.
• Components...
This opens the component properties panel. See "Components" on page 112.
• Binary Interaction Coefficients...
This opens the binaries panel. See "Binary Interaction Coefficients" on page
115.
• Volume Shifts...
See "Volume shifts" on page 115.
• Thermal Properties...
See "Thermal properties" on page 115.
• LBC Viscosity Coefficients...
See "LBC Viscosity Coefficients" on page 116
• Split
This opens the sub-menu of options for splitting fluid components.
See "Splitting components" on page 116.
• Constant Mole Fraction...
• Whitson...
• Multi-feed...
• PNA Distribution
• Group...
• Samples
This opens the sub-menu of sample entry and editing options.
• Names...
See "Sample names" on page 118.
• Compositions...
See "Sample compositions" on page 119.
• Salinities...
See "Sample salinities" on page 119.
• Mix...
See "Mixing samples" on page 119.
• Properties Estimation (FPE)...
See "Fluid Properties Estimation" on page 35.
• Experiments...

Edit
See "Defining Experiments" on page 128.
• Observations...
See "Defining Observations" on page 134.
View
The View menu provides facilities for plotting and reporting.
• To open this menu select View from the main PVTi window.
The View menu has the following options:
• Samples
This option opens a sub-menu containing sample plot types.
• Phase plot
See "Sample phase plot" on page 121.
• Fingerprint plot
See "Sample fingerprint plot" on page 120.
• Ternary plot
See "Sample ternary plot" on page 122.
• Observations...
This allows you to plot an observations against calculated values, or any calculated
values where corresponding observations do not exist.
• Library
This option allows you to view the internal PVTi library. See "Library" on page 105.
Library
The properties of library components are preset by the program.
To display the current list of library components select View | Library...
Table 6.1 List of library components
Mnemonic Name Mnemonic Name

H2 O Water N2 Nitrogen
H2 Hydrogen H2 S Hydrogen Sulfide
CO 2 Carbon Dioxide CO Carbon Monoxide
C1 Methane C2 Ethane
C3 Propane C4 Butane
iC 4 Iso-Butane nC 4 Normal Butane
C5 Pentane iC 5 Iso-Pentane

View
Table 6.1 List of library components (Continued)
Mnemonic Name Mnemonic Name

nC 5 Normal Pentane C6 Hexanes
C6 H6 Benzene C7 H8 Toluene
C7 Heptanes C8 Octanes
C9 Nonanes C 10 Decanes
C 11 Undecanes C 12 Dodecanes
C 13 Tridecanes C 14 Tetradecanes
C 15 Pentadecanes C 16 Hexadecanes
C 17 Heptadecanes C 18 Octadecanes
C 19 Nonadecanes C 20 Eicosanes
C 21 C21’s C 22 C22’s
C 23 C23’s C 24 C24’s
C 25 C25’s C 26 C26’s
C 27 C27’s C 28 C28’s
C 29 C29’s C 30 C30’s
C 31 C31’s C 32 C32’s
C 33 C33’s C 34 C34’s
C 35 C35’s C 36 C36’s
C 37 C37’s C 38 C38’s
C 39 C39’s C 40 C40’s
C 41 C41’s C 42 C42’s
C 43 C43’s C 44 C44’s
C 45 C45’s
Note For components C 6 to C 45 , the properties stored in the internal library

correspond to the “grouped” properties of Single Carbon Number Groups
(SCN), [Ref. 5].

View
Obvious candidates for the pseudoisation of components for use in large regressions
or compositional simulation are iso-butane and normal butane, and iso-pentane and
normal pentane, into single butane and pentane components. A study of many PVT
reports [Ref. 19] has shown that the typical ratios of iC 4 : nC 4 , iC 5 : nC 5 are 0.67:0.33
and 0.60:0.40 respectively.
The library also contains two other components, with the mnemonics C 4 and C 5 ,
whose properties are mole-weighted averages of the respective iso and normal
component properties.
Run
The Run menu provides simulation and regression facilities.
The following options are available:
• Check Fluid System
This provides a consistency check of the current fluid, the results of which are
posted to the log window. If there are a lot of fluid errors, the results are also
displayed in a text window.
• Save As Samples
If this option is turned on, any samples created by an experiment can be saved as
additional project samples.
• Simulate
This simulates all active experiments and then display the simulation results in a
text editor. PVTi has intelligent simulation, which means that the results of the last
simulation run are stored, and if no change has been made to the experimental data
the simulation run is not repeated, the results from the previous run being used.
This keeps the time spent running simulations to a minimum.
• Regression...
This opens the Regression panel. See "Regression in PVTi" on page 137.
Utilities
The Utilities menu option provides access to miscellaneous information relating to the
project and program set-up.
• Units...
See "Units..." on page 155.
• Standard Conditions...
See "Standard conditions..." on page 156.
• Program
This opens the sub-menu of program configuration options.
• Options...
This opens the Options panel which mimics the OPTIONS keyword in the PVI
file. See "Program options" on page 156.
• Debug...
See "Debug..." on page 161.

Run
• Text Editor
This opens the text editor used for displaying simulation results, etc. It can be used
to view any ASCII file.
Graph
The Graph menu provides options related to the plotting of graphs.
• Add New Graph...
Adds a new graph to the existing plot windows.
• Superimpose
When the superimpose option is switched on, indicated by a tick next to the menu
option, subsequent graphs are superimposed on the current main graph.
• Tabulate...
This option creates a table showing the values plotted in the current main graph.
• Remove All
This option deletes all graphs from the window.

Graph
The fluid model
Displaying the SYSTEM section from a PVI file
Displays a RUNSPEC/SYSTEM section present in the current PVI file.
1 To display PVI data, select PVTi: File | View PVI Section | SYSTEM.
Reading the SYSTEM section from a PVI or DATA File

Reads data from a PVI or DATA file. You can use this option to load the equation of
state, viscosity options and hydrocarbon system description from a PVTi PVI file or an
ECLIPSE Compositional (E300) DATA file.
Hint You can load the first two sections of a PVI file as a system specification, rather
than using menu options. Additionally, you can choose to echo the contents of
the whole PVI file to the current print file, PVP.
Reading the PVT section from an E300 DATA file

1 PVTi: File | Import | ECLIPSE Compositional (*.DATA) and select the appropriate
DATA file.
PVTi searches for the required file and, if found, reads it looking for the number of
EoS and Equilibration regions in the ECLIPSE model. The number of reservoir EoS
regions is defined by the ninth entry of the TABDIMS keyword and the number of
Equilibration regions is defined by the first entry of the EQLDIMS keyword. If the
ECLIPSE model has just one of each region type then the program simply reads in
the data. However if multiple EoS or Equilibration regions are found then the
program displays a prompt specifying the numbers of each region found. You are
asked to specify which EoS and/or Equilibration region they wish to read in.
2 Select the number of the EoS and /or Equilibration region you wish to load.
Note EoS regions each have an EoS model defined within them that is an EoS plus a
list of critical properties defined for each component. An Equilibration region
is a group of cells where the initial pressure and saturation is defined. PVTi
needs to know which Equilibration region to read in if there are any
composition versus depth (specified by the ZMFVD or COMPVD keywords)
tables in the ECLIPSE file. There is one table for each of the Equilibration
regions. By specifying which Equilibration region to use this tells PVTi which
table to read in.
Reading the SYSTEM section from a PVI file

1 PVTi: File | Import PVI Section | SYSTEM and select the appropriate PVI file.
PVTi searches for the required file and, if found, reads it looking for all occurrences
of the required section. If there are no RUNSPEC or SYSTEM sections in the file then
no further action is required. However, if one of more sections of the required type
are found in the file, you must select which, if any, are required. The program
displays a prompt specifying the number of sections found.

System
2 Select the section you wish to load.
Note If more than one section is found, the program offers the last section as the
default, although you can read any of the sections.
Hint If you are uncertain as to the contents of the selected section, use File | View
PVI Section to display the section to the screen.
The syntax of the external file is similar to that of ECLIPSE. The data file is free format,
except for keywords, which must start in column 1. For further information on the
keywords see "PVTi keywords" on page 177. An example of such a file for a trivial two-
component CO 2 -isoButane system is as follows:
---
---- RUNSPEC section: specific number of components and the EoS
----
RUNSPEC
NCOMPS
2 /
EOS
PR /
----
---- SYSTEM section: define hydrocarbon properties and composition
----
SYSTEM
---- Unit conventions
UNITS
METRIC ABSOL FRACTION /
DEGREES
KELVIN /
---- Component names (library defaults)
LNAMES
CO2 IC4 /
---- Overwrite default omega values by component
OMEGAA
0.4572 0.4572 /
OMEGAB
0.0778 0.0778 /
---- Initial sample composition
ZI
0.6 0.4 /
---- Binary Interaction Coefficients
BIC
2 1 1 0.13 /
/
----
---- This defines the fluid, EoS etc.
---- COMB, SIMULATE, REGRESS, BLACKOIL sections may now follow
---- See examples in Appendix C
----
END

System
Note the following points:
• Any characters following ---- are taken as comments. The data is free format,
apart from keywords which should start in column 1. You can split data over lines
as required. The forward slash (/) characters terminate data for a keyword.
• You can specify repeat counts for any item. For example 3 * 1.0 implies three values
of 1.0. You can enter defaults by specifying a repeat count alone, such as 1*, or by
the early termination of a data list with a forward slash (/).
• You may enclose character data such as component or experiment names in
quotes. This is optional and is only strictly required when the name contains
embedded spaces or non-alphanumeric characters.
Equation of State
Hint The default Equation of State is the Peng-Robinson three-Parameter equation.

This is suitable for most requirements, so generally you do not need to set the
equation of state.
This panel allows you to choose one of five main equations of state, to specify the
required viscosity correlation, and to decide whether or not to activate editing of
specific heat capacities.
The equations of state are described in "Equation of state" on page 330; the viscosity
options are described in "Viscosity correlations" on page 345 and in [Ref. 5], [Ref. 7] &
[Ref. 42].
The available equations of state are:
• Peng-Robinson
• Soave-Redlich-Kwong
• Redlich-Kwong
• Zudkevitch-Joffe
• Schmidt-Wenzel
Choosing the Equation of State

1 PVTi: Edit | Fluid Model | Equation of State...
This opens the Equation of State and Viscosity panel, which gives you radio
buttons for selecting one EoS from the following list:
• PR: 2-Parameter Peng-Robinson
• SRK: 2-Parameter Soave-Redlich-Kwong
• RK: Redlich-Kwong
• ZJ: Zudkevitch-Joffe
• PR3: 3-Parameter Peng-Robinson
• SRK3: 3-Parameter Soave-Redlich-Kwong
• SW: Schmidt-Wenzel.

System
2 Select the appropriate equation of state.
If you select either of the Peng-Robinson equations or the Schmidt-Wenzel
equation, you must also select whether you wish to use the correction to the
dependence of the Ω a upon acentric factor. The default is the modified (third-
order in ω ) Peng-Robinson form.
• Check the box for Yes, or leave it unchecked for No, as appropriate (see
"Equation of state" on page 330).
Three-parameter extension of the EoS

The three-parameter extensions of the EOS are:
• PR3 - Peneloux et al. three-parameter EoS
• SW - Schmidt-Wenzel EoS (implemented as a modified PR3)
• SRK3 - Peneloux et al. three-parameter EoS.
The PR3 EoS is the default setting.
Hint It has been our experience that the Peneloux et al. three-parameter equations
of state, PR3 and SRK3, generally give much better predictions of liquid
properties and saturations.
They also allow you an additional set of regression parameters, namely the
component volume shifts, making for an easier match to measured data.
Viscosity correlations
The Lohrenz-Bray-Clark, Pedersen and Aasberg-Petersen viscosity models are
available. Select the appropriate viscosity model.
Note You can re-select the equation of state or viscosity correlation at any stage.
However, the default EoS parameters for each component are dependent upon
the EoS, and the program re-initializes these if you change the EoS.
Alternatively you can manually reset the parameters to the default values at
any time.
Components
PVTi: Edit | Fluid Model | Components...
Fluid model components

This option allows you to enter component names and properties. Use this option to
enter new fluid components. You enter a mnemonic and a type, which determines how
the program interprets the component.
1 Select the Names folder.
2 To enter a component, click in an empty index field.
3 Enter the mnemonic for the component and select its type.
See "Component types" on page 113.

System
4 Click on Apply.
5 The other folders now have information in them:
• Complete shows all the properties of all components
• Library shows the properties that were retrieved from the internal library
• User shows user-defined properties
• Characterization allows definition of fluid-model properties by
characterization.
Component types
Library
The PVTi program checks this against the internal library of names. If this exists in the
internal library, it adopts the internal description. If it does not recognize the
mnemonic from amongst the set described in the previous section, you must re-enter
the mnemonic name or respecify the component as a Char or User type, see below.
User
This option allows you to define components. Enter the required properties into the
panel: critical pressure and temperature, acentric factors, etc.
You should enter the components in order of increasing molecular weight, and non-
hydrocarbons before hydrocarbons:
Non-Hydrocarbons
• H2
• H2 O
• CO
• N2
• H2 S
• CO 2
Hydrocarbons
• C1
• C2
• CN+
Hint By selecting PVTi: Run | Check fluid system the fluid is re-ordered into
increasing mole weights.
PVTi allows you to input a user component even if you know only the critical
temperature and pressure. It calculates the other properties as follows:

System
• T b from T c and P c
Riazi-Daubert. For further information see [Ref. 30] and [EQ 8.11], [EQ 8.12].
• Sg from T c and T b
• Mw from T b , T b and Sg
• ω from P c , T b and T c
Edmister. For further information see [Ref. 30] & [EQ 8.10].
• P from
Macleod and Sugden. For further information see [Ref. 14].
• Vc & Zc
Riazi-Dubert. For further information see [Ref. 12].
Use the Update button to calculate the other properties of the component.
Characterization
If you give a characterization, you must generally specify at least two out of the
following (these are specified in the Characterization folder):
• molecular weight M w ,
• specific gravity S g ,
• normal boiling point temperature T b ,
• Watson characterization factor K w ,
• reference temperature K
Hint If you have more than two of the set M w , S g , T b and K w , we recommend that
you enter the best two first, as the order of entry decides which pair the
program selects. For example, if you enter M w , T b and K w then the program
uses M w and T b .
Note It is possible to perform a characterization by entering just the molecular

weight, whereupon the program estimates the specific gravity from a look-up
of Single Carbon Number (SCN) groups.
You can choose from the following correlations for estimating the physical properties
and acentric factors:
Critical properties
• Kesler-Lee. See [Ref. 10].
• Cavett. See [Ref. 11].
• Riazi-Daubert. See [Ref. 12].
• Winn. See [Ref. 44].

System
• Pedersen. See [Ref. 44], [Ref. 42] and [Ref. 46].
Acentric factors
• Kesler-Lee. See [Ref. 10].
• Edmister. See [Ref. 14].
• Thomassen. See [Ref. 30].
• Pedersen. See [Ref. 44], [Ref. 42] and [Ref. 46].
Note When reading in a file the critical volumes (Vc) and critical Z factors (Zc) for
each component must satisfy the relation PcVc=ZcRTc (where Tc, Pc, R are the
critical temperatures, critical pressures and universal gas constant
respectively). If this is not the case then PVTi will alter the values of the
relevant critical Z-factors in order that this relation is satisfied.
Binary Interaction Coefficients

This option enables you to enter the Binary Interaction Coefficients (BICs) for each
component.
1 PVTi: Edit | Fluid Model | Binary Interaction Coefficients...
This displays the Binary Interaction Coefficients panel.
2 Enter the Binary Interaction Coefficients for each component.
3 Alter the Cheuh-Prausnitz-A coefficient as required.
4 Click on Reset to return the interaction coefficients to default values.
Volume shifts
Note Volume shifts are only available if you use a three-parameter Equation of State.
Use this option to enter the dimensionless volume shifts. The actual volume shifts in
the equation of state are displayed beside them.
1 PVTi: Edit | Fluid Model | Volume Shifts... .
This displays the Volume Shifts and Thermal Expansion Coefficient panel.
2 Enter the volume shifts for the required components.
3 Click on OK.
Note If the "Temperature dependence for volume shifts" on page 158 option is set
then you can enter a value for THERMX, the thermal expansion coefficient.
Thermal properties
Note You can only use this option if the program option "Specify/Calculate density
and molar volume units" on page 158 is switched on. It can be switched on in
the Equation of State panel (see "Equation of State" on page 111).

System
Specific heat capacity coefficients and calorific values for each component are the
thermal properties used in PVTi.
1 PVTi: Edit | Fluid Model | Thermal properties....
This opens the Thermal Properties panel.
2 Amend the thermal properties for the components, as required.
3 Click on OK.
LBC Viscosity Coefficients
Note LBC Viscosity coefficients are only available if you are using the LBC Viscosity
model.
Use the option to view or edit the LBC viscosity coefficients.

1 PVTi: Edit | Fluid Model | LBC Viscosity Coefficients....
This opens the LBC Viscosity Coefficients panel.
2 View or amend the coefficients as required.
3 Click on OK.
Splitting components
This menu allows for the automatic splitting of the plus fraction into a required
number of sub-fractions for subsequent use in a large regression or for output to a
compositional simulator such as the one in ECLIPSE.
Splitting is also used to accommodate different plus-fraction properties for different
fluid samples. This process is often known as a multi-feed split.
This option allows you to input data for splitting the plus fraction. There are three
methods available from this option for splitting the plus fraction, which must be the
last component:
• Constant Mole Fraction splitting (CMF)
• Whitson
• Multi-feed split or Semi-Continuous Thermodynamic (SCT) splitting
1 To choose the splitting method, select PVTi: Edit | Fluid Model | Split and select the
splitting option.
Constant Mole Fraction (CMF)

1 Specify the number of pseudo-components you require. The default is N frac = 3 .
2 Give the specific gravity and required sub-fraction split.

By default, the program estimates the specific gravity of the plus fraction from the
reference density, if one was given, and uses a constant mole fraction split of
1 ⁄ N frac .
3 Specify the Whitson Alpha Factor and the Whitson ETA factor, as required.
4 Enter the Critical Props. Correlation and the Acentric Props. Correlation.
5 Give the compositions of the pseudo-components.

System
6 Click on OK.
Whitson or modified Whitson (Whitson)

7 Specify the first single carbon number (SCN) group to be included in the plus
fraction split.
For example, enter 7 if plus fraction is C 7+ .
8 Give the molecular weight, specific gravity and the mole fraction of the plus
fraction.
9 Enter the number of pseudo-components to be used after the regrouping of the
Whitson split.
For example, N MCN = 3 .
10 Specify the Critical Props Correlation and the Acentric Props. Correlation.
11 Select the grouping technique.
12 Choose whether you wish to plot a fingerprint of the Whitson split fractions.
13 Click on OK.
Multi-feed Split (also called semi-continuous thermodynamic (SCT) split)

1 Specify the number of pseudo-components for the split.
This value must be between two and five. The default is two.
2 Confirm the default minimum mole weight in the plus fraction (Whitson η -
parameter) or edit the data as required.
3 Confirm the default mole weight of the heaviest pseudo-component or edit the
data as required.
The default setting is twice the plus fraction mole weight.
4 Set the Critical Props. and Acentric Props. correlations.
5 Specify the group and the molecular weight of the sample’s plus fraction.
6 Amend the default names for the new components, if required.
The default names are FRC1, FRC2, etc.
Note Note that splitting is not necessarily the opposite of grouping. Splitting the
plus fraction into two or more pseudo-components, followed by a re-grouping
of those pseudo-components back into a single plus fraction, generally results
in a different set of critical properties, etc., from those originally possessed by
the plus fraction.
PNA Distribution
This splits all components heavier than the library C6 component into paraffinic (P),
naphthalenic (N), and aromatic (A) components. This is done according to the method
outlined in "The PNA distribution of heavy components" on page 416. The critical
properties assigned to the PNA components are those described in "Critical properties
of PNA species" on page 417.

System
Group
This menu allows for the automatic grouping of sub-fractions for subsequent use in a
large regression or for output to a compositional simulator such as the one in ECLIPSE.
This option allows you to choose components to group and perform the grouping
operation. The default scheme for grouping is to group to the default sample ZI using
the mole fraction weighting to group components. Other schemes of grouping include
grouping by molecular weight and by mixing rule, see [Ref. 45]. Also the sample to
group to can be changed to any in the defined set, or to an average of all samples.
To group components, select PVTi: Edit | Fluid Model | Group...
This displays the current component system, each component having an associated
index. The first time you enter this option, all these indices are set to 0, indicating that
they do not belong to any group.
1 To create a new pseudo component, give a new index of greater than zero to two
or more components.
2 Select the Grouping Technique.
3 Give the group or pseudo-component a new component mnemonic, if required.
Hint You can perform several groupings from the same original component
description by specifying the new components with ascending indices, 1, 2,
etc.
4 Click on the Update button to automatically display any of these new component
names.
5 Click on the OK button to create the groups.
Note Note that splitting is not necessarily the opposite of grouping. That is, splitting
the plus fraction into two or more pseudo-components, followed by a re-
grouping of those pseudo-components back into a single plus fraction,
generally results in a different set of critical properties, etc., from those
originally possessed by the plus fraction.
Defining Samples
Sample names
• PVTi: Edit | Samples | Names...
Defines sample names. Use this option to enter mnemonics for each component. You
can enter more than one sample for later use; to do this, reference each sample by its
mnemonic, of up to 8 characters.
Note Note that the mnemonic for the default sample is “ZI”, for “z initial”.
For alternative samples, you may specify a line of text to give additional information.
For example: from different depths in the hydrocarbon column, a “saved” calculated
composition from a simulation, etc.

System
Sample compositions
• PVTi: Edit | Samples | Compositions....
Enter the compositions for each defined sample. PVTi ensures that they all add up to
unity. If a sample does not add up to unity, a message appears asking whether or not
the program should redistribute the difference across the components.
Sample salinities
• PVTi: Edit | Samples | Salinities...
This option allows you to enter sample salinities. If you have entered H 2 O as a
component then use this option to add the salinity of each sample.
Note This information is used in the MFLASH experiment in the "Simulation using
PVTi" on page 128.
Mixing samples
This option allows you to form a new sample by mixing any two existing samples. You
can enter the amount of each sample to mix either as the mole fraction of the second
sample in the resulting mixture, or as a volume of gas of the second sample as a ratio
to the volume of the first sample at its P sat or other pressure at the specified mix
temperature. The latter case is useful when considering lean gas injections into an oil.
The program produces the required mix provided that:
• The two samples are different.
• The amount of the second sample to mix is greater than zero.
• The number of samples does not exceed the maximum allowed (50).
• The name of the new sample is unique in the set.
1 To mix samples, select PVTi: Edit | Samples | Mix...
This activates the Mix Samples panel.
2 Select the Mixing Type.
3 Choose the fluid samples you wish to mix.
4 Enter the new sample name.
5 Enter the temperature with its units, and the mole fraction.
6 If you are mixing by GOR, give the GOR and the pressure for GOR oil volume
calculation
7 Click on OK.
If the sample is mixed by recombination, the GOR is taken as the stock tank GOR,
the conditions are separator conditions and the mixture is created such that the
stock tank GOR matches the required value.

System
Viewing samples
Sample fingerprint plot
This option allows you to generate fingerprint plots. This consists of plotting the
logarithm of the component mole fractions against the component molecular weights.
Hint Fingerprint plots give an idea of the nature, that is condensate or volatile oil,
of a given fluid sample. Providing a reasonable split of the Heptanes plus is
available, then a condensate typically has straight line or down-turning slope
proceeding towards the heavier fractions, whilst a volatile oil has an up-
turning slope as it usually contains more heavy fractions.
1 To generate fingerprint plots, select PVTi: View | Samples | Fingerprint Plot

This activates the Fingerprint Plot panel, which enables you to select the sample
you require to be used for the plot from a drop-down panel.
2 Select the sample you wish to plot and click on Apply. An example of a plot is
shown in Figure 6.2.
Figure 6.2 Fingerprint Plot

System
Sample phase plot
This option allows you to generate phase plots. This uses the equation of state model
with the current fluid description to obtain the bubble point and dew point lines.
Where the two lines meet is the critical point, at T = T c , p = p c . As part of the
calculation process, an explicit calculation is made of the position of the critical point.
You can choose how many quality lines (lines of constant vapor mole fraction) are
required on the plot; this can vary between 0 and 9 (that is, 10%, 20%,..., 90%).
1 To generate phase plots, select PVTi: View | Samples | Phase Plot
This activates the Phase Plot and Quality Lines panel.
2 Select the sample you wish to plot.
3 Enter the number of quality lines (from 1 to 9).
4 Decide whether or not to plot the Hydrate formation line.
5 Click on OK.
Hint A default phase plot (with one quality line) can be generated by simply
dragging a sample name from the Data Tree and dropping it into the Main Plot
Window.
Note If depletion experiments or separators exist, they are plotted onto the phase
plot too.
Figure 6.3 Phase plot

System
Sample ternary plot
This option allows you to create a ternary plot for a particular sample. The ternary plot
panel allows you to set: the sample to be plotted; the temperature and pressure for the
plot; and the grouping of the fluid components so as to create three components for the
ternary plot.
1 To generate a ternary plot, select PVTi: View | Sample | Ternary Plot
2 Select the fluid sample for plotting
3 Enter a temperature and pressure.
4 Select the component groupings and the names of the grouped components.
Hint The default component groupings are: C1 and the non-hydrocarbons, C2-C6
and C7 and heavier hydrocarbons. This is typically the best choice, so you
should only need to change the groupings from the default in special cases.
Figure 6.4 Ternary Plot

System
COMB - Compositional Material Balance
Introduction
Material balance checks provide an important tool in analyzing the quality of the data
found in a PVT report. In PVTi, material balance checking is provided for Constant
Volume Depletion (CVD) experiments, Separators by recombination, and the
calculation of liquid and vapor recovery.
Caution It is important to always examine PVT laboratory data for material balance
errors. If the reported observations contain serious errors, these will be
reflected in the fitted equation of state model.
It is generally the case that for gas condensate and volatile oil samples, a constant
volume depletion experiment (CVD) is performed as part of a laboratory analysis.
Using material balance considerations, it is possible to calculate liquid compositions
and hence K -values, molar masses, densities, etc. This can be very useful for two
reasons:
1 With appropriate separator data, estimates of oil and gas recovery can be
performed without recourse to an equation of state model.
2 The consistency checks provide a measure of the quality of laboratory data and
consequently its value or otherwise in any subsequent regression analysis.
Accessing material balance checks

The Material Balance panel can be found by selecting the experiment in the project
tree-view, right-clicking on it and selecting Material Balance... from the pop-up menu
that appears.
A technical outline of the methodology used in the Material Balance panels can be
found in "Consistency checks and correlations" on page 319.
There are three different Material Balance panels: one for depletion (CVD), one for
separators, and one for Constant Composition Expansion (CCE). Each is described
below.
Material balance for Constant Volume Depletion (CVD)

Experiments
Right-click on the experiment in the project-tree and select Material Balance... from the
pop-up menu.
There are three buttons on this panel
• Report
• Plot
This gives access to the full range of material balance diagnostic plots. For more
details on these plots see "Consistency checks and correlations" on page 319.

COMB Module
• Modify
Report
This option performs the material balance calculation and produces a report on any
current problems with the fluid.
Figure 6.5 Main display after performing material balance
Plot
Use this option to perform plots.
The set of plots available from this option are:
• Vapor composition (input)
• Liquid composition
• K -values (1): log( K ) versus pressure
• Produced moles
• Vapor moles left in cell

COMB Module
• Liquid moles left in cell
• Produced mass
• Vapor compositions versus pressure
• Liquid compositions versus pressure
• Liquid volume in cell
• K -values (2): Hoffmann plot
• Initial and recovered compositions
• Final stage liquid composition
• Calculated and correlated liquid density
• Input and correlated vapor Z -factor
Figure 6.6 COMB module - vapor versus pressure plot
Modify
Use this option to modify data. Errors in the input data may become evident after you
performed the CVD material balance. You can use this option to rectify these errors.
You may consider changing the saturation pressure liquid saturations, Z -factors, etc.
Hint You can avoid calculated liquid mole fractions remaining in the cell by
changing the composition of the feed stream (well stream) or one or more of
the removed gas streams.

COMB Module
Negative calculated liquid moles
Negative calculated liquid moles for a given component across a wide range of the
sampled pressure stages probably indicate errors in the wellstream composition
(which is usually a calculated recombination sample).
Errors at a lesser number of stages or at just one or two specific stages probably indicate
a measurement error on a given removed gas stream.
In either case, you can modify the well stream or removed gas stream compositions by
increasing (positive % or absolute value) or decreasing (negative % or absolute value)
one or more component compositions. The program automatically re-scales the
remaining (unmodified) compositions in the stream, according to their initial mole
fraction, so that the total mole fraction for the stream sums back to 100%.
If you require any of the compositions to remain fixed prior to the material balance
calculation, type any character in the field immediately preceding the Modifier field for
that composition.
Separator recombination
Use this panel to directly test the quality of recombinations. To open this panel,
right-click on the separator experiment in the project-tree and select Material Balance...
from the pop-up menu.
Report
Use this option to test recombinations. It generates the calculated and Standing's K -
values and provides a report on the results.
Testing combinations
If sufficient volumetrics data has been entered in the previous item the recombined
sample is calculated and output to the report and PVP file.
The output details the calculated feed to either separator using the given values of
GOR, oil density and liquid and vapor compositions from each separator stage and is
printed together with the given feed composition for comparison.
Plot
Use this option to plot the recombination results. It produces a Hoffmann-Crump-
Hocott plot, which gives a measure of the quality of the separator data.
Constant Composition Expansion (CCE) recovery

calculations
Use this option to make estimates of the recovery of vapor and liquid from an initial
reservoir pressure in excess of the saturation pressure defined in the CVD experiment.
To open this panel, right-click on the CCE experiment and select Material Balance...
from the pop-up menu.
Note This option is only available if there are valid depletion and separators
experiments in the project, in addition to the CCE experiment.

COMB Module
Rel. Perm.
Use this option to define liquid production. If oil is to be ‘produced’ in the recovery
calculations (using the method of Reudelhuber Hinds, [Ref. 37]), you must enter two
points on the plot of relative permeability versus total liquid saturation.
Report
Use this option to perform recovery calculations. It allows you to estimate recoveries
of vapor and liquid.
Plot
Use this option to produce a plot of the gas and oil recovered as a function of pressure.

COMB Module
Simulation using PVTi
Introduction
PVTi allows you to perform experiments on your defined hydrocarbon system. The
experiments available include:
• Flash calculation
• Bubble and dew point determination
• Simulation of laboratory experiments such as constant composition expansion,
constant volume depletion, and differential liberation
• Swelling tests
• Separators
• Variation of composition and pressure with depth
• Vaporization tests
• multiphase flash.
• Wax appearance temperature
• Asphaltene appearance pressure
Note You can enter any experimentally determined observations available to

compare with the output produced by the Simulate module, apart from the
multiphase flash experiment.
Defining Experiments
1 PVTi | Edit | Experiments...
This opens the Define Simulate Experiments panel.
2 To add a new experiment select it on the Add menu. To edit an existing experiment,
select its name from the drop-down list, which appears in the panel, and click OK.

Simulate
3 Once an experiment has been created for definition or selected for editing, you will
see a customizable form that can be tailored to suit your data. The information is
entered in a number of folders.
• In the first folder, General, you can select various single-valued pieces of
information for entry. The required information is automatically selected and
cannot be deselected, so an information line informs that this data is required.
• The second folder, Observations, shows a table where you can customize
column headings to match your observation data. If you are editing an existing
experiment, the observations in the currently defined set are already shown as
column headings. In this way you can match the data entry panel to your own
data-set. If you forget to enter a required quantity (for example Pressure in a
Differential Liberation experiment - see "Differential liberation" on page 130)
PVTi warns you and does not create the customized entry form.
• The third folder, Components, allows you to determine whether you enter
component-base data or not. Typical options here are for Liquid Mole
Fractions, Vapor Mole Fractions or K-values.
• The fourth folder, Other, is used for miscellaneous observations that do not fit
any of the other categories. Currently this folder is only used by the Constant
Volume Depletion experiment for the Final Liquid Mole Fraction. When other
experiments are being entered, this folder does not appear.
• Click Apply.
4 A customized form is now created, with the same folders as described above. Now
the folders contain data-entry fields and tables for observations. Once the data
have been entered, click on Apply to submit the data and create or edit the
experiment.
5 Finally, Close becomes active and can be used to close the panel
Data requirements for the experiments

Flash calculation
For this experiment you must define the pressure and temperature of the flash.
The program performs a stability test and establishes the number of phases present
prior to the flash calculation.
Note The gas-oil ratio reported by the calculation is defined as gas volume at
standard conditions divided by liquid volume at flash conditions. The gas
volume is obtained using a Z -factor of unity.
Bubble point pressure

For this experiment you must enter the temperature at which the bubble point is
required.
Note If the temperature is such that no bubble point can be found (above the critical
temperature) the program returns a warning message.

Simulate
Dew point pressure
For this experiment you must supply the temperature and choose between normal or
retrograde dew points. The default dew point is retrograde.
Note If the temperature is such that no dew point can be found (above the critical
temperature) the program returns a warning message.
Constant composition expansion

For this experiment you must specify a temperature and a series of pressures.
Additionally you must specify whether the fluid is oil or gas. You do not need to give
a value at saturation pressure.
Hint You can apply this experiment to a liquid (bubble point) or vapor (dew point)
system. The program tests for both possibilities. It is also possible to perform
a constant composition expansion on a true one-phase system (SIN), such as
an (dry) injection gas above its cricondentherm.
Note When obtaining relative volumes the program uses saturation volume as a
normalisation volume, if one exists, or the volume at the highest pressure, if
not.
Constant volume depletion

For this experiment you must specify a temperature and a series of pressures. You do
not need to give a value at saturation pressure.
Hint You can apply this experiment to a liquid (bubble point) or vapor (dew point)
system. The program tests for both possibilities. It is not, however, possible to
apply this experiment to samples that are above the cricondentherm.
Note The relative volume reported by the program is the fraction of the cell filled
with liquid at the end of the constant volume step, that is after the original
volume has been restored by removing vapor.
Differential liberation
For this experiment you must specify a temperature and a series of pressures. You do
not need to give a value at the bubble point. PVTi provides this pressure point.
The program also provides automatically the last step in the differential liberation
process, the reduction to standard conditions. However, the program does not provide
the pressure point at standard pressure (usually 14.7psia) and at reservoir temperature
and the user must enter this for the final stage.
Note You may only apply this experiment to a liquid (bubble point) system.

Simulate
Hint The relative volume reported by the program is the ratio of the oil volume at
each step to the oil volume at the final (standard conditions) step.
Note There are alternative definitions of the GOR and the relative oil volume
available using the program options "Definition of GOR in Diff. Lib." on page
159 and "Definition of Oil relative volume in Diff. Lib." on page 160.
Swelling test
A swelling test consists of adding increasing amounts of a lean gas to a reservoir fluid
and determining the swelling of the mixture relative to the original fluid composition.
For this experiment you must specify:
• The nature of the original fluid type, OIL or GAS.
• The composition of the lean gas to be added.
• The reservoir temperature.
• A set of either mole percentages of gas in the mixture or GORs (the volume of gas
at STC, that is 14.7 psi and 60 °F , per volume of oil at original saturation pressure
or other specified pressure).
Separators
Separators consist of a set of connected equilibrium flashes at user prescribed pressures
and temperatures.
• The composition of the feed stream from the defined sample mnemonics.
• A number of stages (up to seven) for which you must give a pressure and
temperature.
Additionally, you must connect the vapor and liquid outputs of each stream to some
subsequent stream. The default routing is to connect the liquid output of stage (j) to
stage (j+1), and to take the vapor output to stock tank conditions (as defined by the
STCOND keyword or by the Standard Conditions menu option under OPTIONS).
Note A stage output can be fed back to any previous stage though not back to the
current stage. No stock tank stage is defined automatically, and whereas it is
customary to quote vapor properties at stock tank conditions, liquid properties
will be quoted at the “final stage” conditions. Therefore, if liquid properties at
stock tank conditions are required, this should be the final (additional) stage
which must be defined by the user. For example if we have a separator with 3
stages with the last stage being stock tank conditions, then a liquid FVF at
stage 1 of the separator will be the volume of liquid divided by the final liquid
volume (stock tank conditions in this case) after flashing the liquid feed of
stage 1 through the remaining 2 stages of the separator chain.

Simulate
Hint The "Definition of Oil relative volume in Diff. Lib." on page 160 program
option allows you to quote GORs as volume of gas at standard conditions per
volume of stock tank oil as opposed to the default calculation of volume of gas
at standard conditions per volume of separator liquid at separator conditions.
The program can calculate oil formation volume factors, that is the volume of reservoir
fluid at initial or bubble point conditions per stock tank volume (SRELV) and, by
separator stage, volume of separator liquid at separator conditions per stock tank oil
volume (ORELV).
To use this option tick the box in the panel or use the FVFREF keyword in batch mode.
Variation of composition and pressure with depth

It is well known that composition varies with depth in a reservoir.
• A reference sample composition (from the currently defined sample mnemonics).
• A reference depth, pressure and temperature for the sample.
• A set of depths above and/or below the reference depth, at which you wish the
program to calculate the composition and pressure.
If during the increment up and/or down, either a genuine gas-oil contact is found or a
transition from gas to oil (or vice-versa) without passing through a contact (a “critical”
transition), then the program reports this depth.
Note The assumptions made in the performance of this experiment, that there are no
asphaltenes and that the reservoir is in thermal, gravitational and diffusive
equilibrium, are probably not achieved in any real reservoir. However, despite
these reservations, this is a useful test of the depth-variation of a particular
fluid.
Vaporization test
This is a somewhat specialised test performed for gas-injection on reservoir fluids, but
in fact it is rather similar to a swelling test.
• The composition of the reservoir fluid and injection fluid from the currently
defined sample mnemonics.
• The reservoir pressure and temperature
• The number of moles of injection gas to be added to the reservoir fluid.
Multiphase flash calculation

The inputs required for the multiphase flash experiment are the same as for the usual
two-phase flash experiment ([Ref. 36]). For this experiment you must define the
pressure and temperature of the flash. The program performs a stability test and
establishes the number of phases present prior to the flash calculation.

Simulate
Note The gas-oil ratio reported by the calculation is defined as gas volume at
standard conditions divided by liquid volume at flash conditions. The gas
volume is obtained using a Z -factor of unity. Note also no facility currently
exists for comparing these against observed values.
Note If the sample you select for the experiment contains water, you should enter
the salinity in the PROPS section.
Saturation pressure calculation

This is essentially the same as the bubble and dew point calculations. For this
experiment you must enter the temperature at which the saturation point is required.
The calculation is particularly useful if you have no a priori information about whether
the saturation point is bubble or dew.
Note In the case of the dew point, program returns the retrograde (highest pressure)
dew point.
Saturation temperature calculation

For this experiment you must specify the pressure. Since there are generally two
saturation temperatures (one from each side of the phase envelope), you must also
specify which solution is required - lower or higher.
Critical point calculation

This is a convenient way of obtaining the critical point of a sample, without generating
a full phase envelope. Enter the sample name for this experiment.
First contact miscibility pressure calculation

This calculation returns the lowest pressure at which the samples are directly miscible,
regardless of the proportions in which they are mixed. The method used to determine
the minimum pressure is described in the paper by Jensen and Michelsen, [Ref. 39]. For
this experiment specify the temperature and the names of the two samples.
Multiple contact miscibility pressure calculation

This calculation determines the lowest pressure at which two samples (one oil, one gas)
are always miscible (regardless of their relative proportions) after repeated contacts
between them, when only one of the samples is affected at each contact. When the
sample affected is the gas, this simulates a one-cell vaporizing drive. If the oil is
affected each time, this mimics a one-cell condensing driveways method used to
determine the minimum pressure. This case is also described in the paper by Jensen
and Michelsen, [Ref. 38]. For this experiment, specify the temperature and the names
of the two samples. Give the drive to simulate.
Multiple contact test

This experiment simulates the multiple contact test where a series of flashes are
performed on mixtures of reservoir oil and injected gas.

Simulate
For this experiment specify:
• The oil and gas samples
• The temperature and pressure of the test
• The drive to simulate (either condensing where the remaining oil is kept after each
flash and contacted with the initial gas sample, or vaporizing where the remaining
gas is kept after each flash and contacted with the original oil sample)
• The fractions of remaining oil/gas to be contacted with the original gas/oil at each
stage.
Hint You can use this experiment in the REGRESS section. For further information
see "Regression in PVTi" on page 137.
Defining Observations
Observations can be defined at the same time as the experiment; see "Defining
Experiments" on page 128.
1 PVTi: Edit | Observations...
This opens the Define Simulate Observations panel.
2 Choose the index of the experiment
3 Select abbreviation for the observation to be entered
4 Enter the data for this observation
5 If you are entering data for regression purposes, give weightings, either individual
or global, for the observation types.
The observations available vary with experiment type, but will be from the following
set:
Table 6.1 Observation data
Abbreviation Observation
Liquid Z-Factor Liquid Z -factor
Vapor Z-Factor vapor Z -factor
Liquid Density Liquid density
Vapor Density Vapor density
Liquid Mol. Wght. Liquid molecular weight
Vapor Mol. Wght Vapor molecular weight
Liquid Visc. Liquid viscosity
Vapor Visc. vapor viscosity
Liquid Sat. Liquid saturation
Vapor Sat. Vapor saturation
Vapor Mol. Frac. Vapor mole fraction
Sat. Pressure Saturation pressure: gas - p dew , oil - p bub
Sat. Temperature Saturation temperature: not currently available

Simulate
Table 6.1 Observation data (Continued)
Abbreviation Observation
Gas-Oil Ratio GOR: SEPS - gas (STC)/oil (stage/STC); DL gas (STC)/oil
(STC/ p sat )
Relative Vol. Relative volume (SWELL = swelling factor)

Total Gas-oil ratio Cumulative separator GOR: (Gas at STC/final stage Oil)
Ternary Plot Ternary Plot (Multi-Contact Test and plotting only)
Mole. wght. plus Mole weight of plus fraction (COMB Mat. Bal.)
Rel. oil sat. Relative oil saturated volume ( B o ( p bub ) in DL)
K Values K -values
Liquid mol. frac. Liquid mole fraction
Vapor mol. frac. vapor mole fraction
Total mol. frac. Total mole fraction
Spec. grav. plus Specific gravity of plus fraction (COMB Mat. Bal.)
Moles Recov. Moles recovered from depletion experiment (CVD,DL)
Liquid mol. vol. Liquid molar volume (specific volume)
Vapor mol. vol. vapor molar volume (specific volume)
Final mol. wght. Mole weight of liquid plus fraction (COMB Mat. Bal.)
liq. plus
Final spec. grav. Specific gravity of liquid plus fraction (COMB Mat. Bal.)
liq. plus
Final liq. mol. Liquid mole fraction of final stage of CVD (COMB Mat. Bal.)
frac.
Total rel. vol. Total (oil and gas) relative volume (DL)
Oil rel. vol. Oil relative volume (DL, SEPS, vapor) (also see RVSAT)
Gas Gravity Gas gravity (Differential Liberation)
Gas FVF Gas formation volume factor (DL)
Gas Vol. Ext. Gas volume extracted (at STC) (DL)
2-phase Z Two phase Z -factor (CVD)
Oil rel. FVF Oil FVF from p init ⁄ p bub to p stock (SEPS)
Note Note that not all the observed data types are available for all experiments.
Running a simulation
Note Simulations are automatically run on creation, so the results are immediately
available.

Simulate
When the simulation is complete the program displays a text module detailing the
success or otherwise of the runs. The PVTi main display window showing the
experiment results will resemble the following:
CCE : Constant Composition Expansion

Soave-Redlich-Kwong (3-Parm) on Z1
Lohrenz-Bray-Clark Viscosity Correlation
Density units are KG/M3
Viscosity units are CPOISE
Surface Tension units are DYNES/CM
Specified temperature Deg K 377.5944
Liq Sat calc. is Vol oil/Vol Fluid at Sat. Vol
------------------- ----------------------- ------------ ------------
Rel Volume Vap Mole Frn Liq Density
Pressure Inserted ----------------------- ------------ ------------
BARSG Point Observed Calculated Calculated Calculated
------------------- ----------------------- ------------ ------------
344.739 0.9189 0.9485 682.4368
310.265 0.9278 0.9565 676.7593
275.791 0.9379 0.9652 670.6133
241.317 0.9492 0.9750 663.9241
199.948 0.9623 0.9882 655.0453
193.053 0.9651 0.9906 653.4620
186.159 0.9681 0.9930 651.8460
179.264 0.9711 0.9955 650.1961
Hint PVTi recognizes which experiment simulations are up to date and then only
performs necessary calculations. This means that to view the simulation
results you should always use PVTi: Run | Simulate.
Hint If you click on an experiment in the sample-tree using the right mouse button,
and select Report... from the drop-down menu, you can see the report for that
experiment on its own.
Plotting simulation results

In addition to the simulation results tables, the results of simulations can be plotted.
There are two ways to do this.
Firstly, to view the comparison between the simulated results and the observations,
simply drag the appropriate observation from the Data Tree and drop it into the Main
Plot Space.
Secondly, you can use the observation editor to plot any simulated quantity, not just
those for which there are observations. For information on the observation editor see
"Defining Observations" on page 134.

Simulate
Regression in PVTi
Introduction
Performing a regression
To perform a regression you must specify:
• The experiments to be used in the regression.
You can choose from the experiments mentioned in "Defining Experiments" on
page 128.
• The weighting for the observations associated with those experiments.
You can use most of the observations given in a laboratory experiment as
observations to match against predicted data.
• The Equation of State parameters you wish to vary to match predicted to observed
quantities.
Most of the Equation of State parameters are available as regression variables.
Note The time taken for the regression operation rises rapidly with the number of
variables chosen, and the use of the minimum possible set is suggested. That
said, any combination of critical point data, Ω a and Ω b values, acentric factors,
binary interaction coefficients, δ ij and volume shift parameters (if the PR3 or
SRK3 forms for the Equation of State are being used) may be chosen to be
modified. There is a maximum total of 50 regression variables.
Regression function
The regression function to be minimized is the normalized root mean square (RMS)
error of predicted experiment results to the given (weighted) observed experiment
results. See "Details Folder" on page 142 for a description of the RMS value used.
Note In order to run the regression, there must be at least as many observations as
chosen regression variables.
Regression Panel
To open the Regression panel select PVTi: Run | Regression...
Use this menu to set up and perform regression. To open this menu, select PVTi: Edit |
Regression
Variables
The Variables section has the following options:

Regress
• Normal (component properties and BICs)
Normal variables are individual fluid component properties and binary
interaction coefficients. For further information see "Defining regression
variables..." on page 138.
• Special (MW of characterized components etc.)
The special variables available depends on the project settings and the fluid model
properties. Typical special variables are the mole weight of the plus fraction or the
Cheuh-Prausnitz coefficient for binary interaction coefficients. See "Setting special
variables" on page 140.
• PVTi Selects
This sets up the regression variables according to the rules given in "Physical
selection of regression parameters" on page 408.
The two buttons are:
• Variables...
Opens a panel specific to the selection of variable types (see above).
• Limits...
Sets limits on regression variables. For further information see "Setting regression
limits" on page 140.
Report
The Report section has the following buttons:
• Regression
Opens the Regression Report panel. See"Regression Report" on page 142.
• Simulation
Opens the simulation report of all experiments.
Regress
The Regress section is for running regressions. The buttons in this section are:
• Run
Perform regression. For further information see "Performing a regression" on page
141.
• Accept
• Reject
Accept or reject the last regression. For further information see "Accepting or reject
regression results" on page 141.
Additional Information
Defining regression variables...
Use this option to define/re-define the set of variables for use in the regression. There
may be two sets of variables available for use in regression, depending on the state of
the "Program options" on page 156 and whether modified Whitson splitting (SCT) has
been used on the plus fraction. The two sets are denoted normal and special.

Regress
• The normal variables are the component dependent ones, that is variables such as
critical properties, acentric factors, etc.
• The special variables are system-wide or multi-component variables such as the
thermal expansion coefficient or the Cheuh-Prausnitz A -factor for binary
interaction coefficients.
Setting normal variables

There are two panels for setting normal variables.
Use the first panel to define the EoS parameters, the parameters for the LBC viscosity
correlation and the volume shifts for the currently defined N c components.
1 Enter integers for the EoS parameters:

• Pc
• Tc
• ω
• Ωa
Hint If you wish to vary a given EoS parameter, say T c , of two or more components
as separate independent quantities, you should give them different values. For
example 1,2,..., etc.
2 Enter integers for the LBC viscosity parameters:

• Ωb
• Vc
• Vc
Hint If you wish to vary the parameters as one or more groups of variables, you
should give the required group members the same integer. This may be
particularly useful when trying to vary V c values to match to viscosity data
using the LBC correlation, for example.
3 Enter integers for the volume shifts:

• PR3
• SRK3
Note All of these data fields can take an integer value, 0,1,2,..., and so on. The default
of zero (or blank/null field) implies that the particular component’s EoS
parameter is not to be used in any subsequent regression.
Use the second panel to define the status of the lower half of the (symmetric, zero-
diagonal) matrix of binary interaction coefficients.

Regress
Note Note that the rules regarding choice of groups for binary interaction
coefficients are slightly different in that groups may be specified down
columns or along rows of the lower half matrix but not both.
Setting special variables

1 Give the following information:
• Plus fraction mole weight.
• Plus fraction skewness.
• Thermal expansion coefficient for volume shifts.
• Pre-multiplying coefficient for Cheuh-Prausnitz BICs.
• Characterization for SCT-splitting.
Note Note that the mole weight and skewness variables apply on a sample-by-
sample basis, therefore there must be the appropriate number of
measurements defined to allow this option to be used.
Setting regression limits

Use this option to change the regression control parameters.
1 Select PVTi: Edit | Regression | Limits...
This opens the Regression Controls panel where you can set the following options.
Maximum number of iterations

The maximum number of successive iterations that the regression uses. Enter the
maximum number of iterations. The default value is 10.
Maximum step limit

The maximum amount if change allowed in the regression vector during successive
iterations. Give the maximum step limit, if required.The default value is 0.100000. For
further information see "Regression" on page 365.
Minimum step limit

The minimum amount of change allowed in the regression vector during successive
iterations. Enter the minimum step limit, if required. The default is 0.000010. For
further information see "Regression" on page 365.
Regression target
The regression target is the size of the objective function at which the regression
terminates as having achieved a match of calculated to observed values. Enter the
regression target. The default value is 0.000001.

Regress
Hint Normally you do not need to change any of these limits, except for the
maximum number of iterations (which might be reduced, especially for larger
problems) and the limits on some variables (for example to prevent violation
of monotonicity relationships).
Note For all variables specified in the regression the program displays the lower and
upper limits.
Variable Lower limit Upper limit

VAR1 VAR1lo VAR1up
VAR2 VAR2lo VAR2up
: : :
: : :
Note For all the equation of state parameters, except the acentric factor and binaries,
these lower and upper limits are scaled (to unity) variables, with default
settings of 0.5 and 1.5. That is, the program allows the variable to
decrease/increase by up to 50% before it terminates the regression.
Running a regression
Use this option to perform regression.
Note This option is only available if you first define a set of experiments,
observations and variables.
Performing a regression
1 PVTi: Run | Regression
The program performs the regression after first checking that there are at least as
many regression points as variables.
Hint The results of regression can be viewed in the Sensitivity Analysis panel. See
"Regression Report" on page 142.
Accepting or reject regression results

Use this option to accept or reject the last regression The program holds the regressed
system in memory. This allows you to examine the plots and experimental output and
decide whether you wish to accept or reject the regression.
Accepting/rejecting a regressed system

1 Examine the plots and experimental output

Regress
2 PVTi: Run | Regression | Accept/Reject
3 Accept or reject the regression as appropriate.
Regression Report
The regression report contains detail of the current fit between the model and the
observations.
Also use this option to display the sensitivity matrix, that is the sensitivity of the
program’s predictions to each of the given regression variables. This panel also gives
detailed descriptions of the current fit and the conditioning of the regression problem.
All the information required to develop a mathematically sound regression problem is
available through this panel. The hints provided in this section help you interpret the
wealth of information contained within the folders of this panel.
Hint In any regression, having a few very sensitive parameters is preferable to

having hundreds of insensitive ones. Always look for parameters that can be
discarded. This is called conditioning the problem - an ill-conditioned problem
is difficult to solve.
Details Folder
The first folder, Details, shows the current fit. The numbers at the top of the folder show
the Total (normalized) RMS fit and the Weighted (normalized) RMS fit.
Hint The RMS values in PVTi are normalized (the difference between the observed
and calculated values is divided by the observed value) so that, for example,
pressures (which could be thousands of psi) are treated similarly to saturations
(which are between zero and one). To calculate the weighted RMS, each
normalized difference is multiplied by the assigned weighting. The default
weight is 1 and so, initially, the two RMS values are the same.
The remainder of this first folder shows the observations, their weighting in the
regression, and the percentage difference between the observed and calculated values.
Modifiers folder
The second folder, Modifiers, shows the selected regression parameters, the minimum
and maximum allowed values for each modifier, and the percentage change made to
the modifier during regression (initially this is zero as no regression has been run).
Sensitivities folder
The Sensitivities folder shows the sensitivity of each observation to changes in each
regression variable.
Hint A large sensitivity indicates that changing that regression variable has a large
effect on the fit to that observation. Likewise, consider discarding regression
variables to which the observations are insensitive, since a large modifier is
needed to obtain a fit and, in general, large modifications lead to unrealistic
fluid models.

Regress
Hessian folder
The Hessian matrix is a good indication of the conditioning of the inversion problem
(regression).
Hint In a well-conditioned problem the leading diagonal of the Hessian matrix is

dominant. Specifically, look at an element on the leading diagonal. If it is larger
than the other values in that row (or column) of the Hessian, then it indicates
that the regression is likely to succeed.
Covariance folder
The Covariance matrix shows the likely scale of variation in the regression variables
that will occur during regression. The larger a value, the less well-determined the value
of the regression variable will be.
Correlation folder
The Correlation matrix is very important as it can indicate links between regression
variables that are not obvious on first inspection. If two variables are strongly
correlated (correlation close to 1), they both move the fit in the same direction; and so
changing one is similar, in effect, to changing the other. If two regression variables are
strongly anti-correlated, changing one has the opposite effect to changing the other.
This latter case can cause a difficulty in regression as the two variables could be
changed an unlimited amount in opposite directions without having a noticeable effect
on the fit.
Hint Look at off-diagonal elements in the correlation matrix. If any are close to 1 or
-1, consider removing one of the two regression variables that are correlated.
This improves the likelihood of a good final fluid model being created.
(If the variables are of the same type, for example if they are both Tcrit, you
could consider combining them into a single regression variable.)

Regress
Exporting keywords
General information
PVTi can be used to generated output for ECLIPSE BlackOil, ECLIPSE Gi option,
ECLIPSE Compositional, ECLIPSE Thermal, VFPi and the API Tracking option in
ECLIPSE BlackOil.
The Export modules are used to produce models for the ECLIPSE simulators (see
"Output for ECLIPSE simulators" on page 372 for background information). You can
generate data files for exporting to ECLIPSE BlackOil , ECLIPSE Compositional,
ECLIPSE Thermal, API Tracking and VFPi from PVTi.
To open this menu, select PVTi: File | Export.
The Export menu has the following options:
• ECLIPSE Compositional Fluid Model...
See section entitled "Export for ECLIPSE Compositional".
• ECLIPSE Thermal Fluid Model...
See section entitled "Export for ECLIPSE Thermal".
• API Tracking option in ECLIPSE BlackOil...
See section entitled "Export for API Tracking option in ECLIPSE BlackOil".
• Oil reservoir...
See section entitled "Export Oil Reservoir".
• Gas reservoir...
See section entitled "Export Gas Reservoir".
• Equilibration...
See section entitled "Export Equilibration"
• Water...
See section entitled "Export Water".
• VFPi
Use this option to generate VFPi tables. For further information see "VFP module"
on page 149.
Export for ECLIPSE Compositional

Description
This panel exports the fluid model as keywords suitable for the ECLIPSE
Compositional PROPS section.
Units
Allows you to export in the ECLIPSE unit set of choice.

Export modules
Sample
If a sample is selected, it will be exported in the ZI keyword.
Hint If you are going to use COMPVD (recommended), you do not want a ZI
keyword, so set the Sample to {none}
Reservoir temperature
This is exported in the RTEMP keyword.
Export for ECLIPSE Thermal

This panel exports the fluid model as keywords suitable for the ECLIPSE Thermal
PROPS section.
Sample
The sample for which the keywords will be exported for.
Number of Flashes to be Performed

This number is used when calculating the coefficients of Crookston’s equation. You an
usually leave this set as the default value of 20.
Max/Min Pressure/Temperature
You are recommended to enter the maximum and minimum values of pressure and
temperature for your reservoir.
Export Crookston Coefficients

If the box is ticked then the coefficients of Crookston’s equation will be exported. If not
then Wilson’s formula is used to calculate K-values.
Units
A detailed explanation of the workflow required for this export option is described in
the "Compositional Data for ECLIPSE Thermal" on page 387. For a technical review of
PVTi’s export facility for ECLIPSE Thermal see "ECLIPSE Thermal Export Module" on
page 423.
Export for API Tracking option in ECLIPSE BlackOil

This panel can be used to export a series of black oil tables suitable for use with the API
Tracking option in ECLIPSE BlackOil.
Set of Fluid Samples

Select which set of fluid samples to export the black oil tables for.
Properties Keyword
This sets the keywords that are exported, for example Live Oil (PVTO) and Dry Gas
(PVDG).

Export modules
Hint All the keywords are described in the "ECLIPSE Reference Manual".
Write Gas Tables for each Sample?

If exporting the PVDG (dry gas) or PVTG (wet gas) tables then you can specify whether
to write out a gas table for each sample. Often a table for each sample is only required
for the oil keywords in ECLIPSE. If No is selected then PVTi exports a gas table for the
sample in the list with the median vapor density at surface conditions.
Plot Results?
You can tell PVTi not to plot the tables if you wish. PVTi only has room for 8 plots so if
there are many samples in the list then it may be useful not to plot the tables.
Write Values to Double Precision?

You can ask for full double precision values if you wish, but the table columns may not
be fully aligned if this option is used.
Separator Experiment
You can select a separator in your project if you wish instead of the default standard
conditions separator (usually 14.7psia and 60F).
Table Generation Method

The algorithm used to generate the black oil tables. See "Output for ECLIPSE
simulators" on page 372.
Units
Export Oil Reservoir

This panel exports keywords from a Differential Liberation experiment (that is, for an
oil reservoir).
The keywords are suitable for the ECLIPSE BlackOil and ECLIPSE Compositional
PROPS section.
Units
Properties Keyword
This sets the keywords that are exported.
Write Values to Double Precision?

You can ask for full double precision values if you wish.

Export modules
Separator experiment
If creating tables for ECLIPSE BlackOil, the output from the Differential Liberation can
be passed through any separator. The default is to use a single stage at standard
conditions (that is, a stock-tank only - normally 14.7 psia, 60F).
Table generation method

See "Output for ECLIPSE simulators" on page 372.
Export Gas Reservoir

This panel exports keywords from a Constant Volume Depletion experiment (that is for
a gas reservoir).
The keywords are suitable for the ECLIPSE BlackOil and ECLIPSE Compositional
PROPS section.
Units
Properties Keyword
This sets the keyword that is exported.
If creating tables for ECLIPSE BlackOil, the output from the Constant Volume
Depletion can be passed through any separator. The default is to use a single stage at
standard conditions (that is, a stock-tank only).
Table generation method

See "Output for ECLIPSE simulators" on page 372.
Injection fluid
For the Gi option only, this provides the injection sample to be used in creating the
pseudo-compositional keywords. See "Pseudo-compositional tables for ECLIPSE
GI option" on page 380.
Export Equilibration
This panel exports keywords from a Composition versus Depth experiment.
The keywords are suitable for the ECLIPSE BlackOil INIT section and ECLIPSE
Compositional PROPS section.
Units
Properties Keyword
This sets the keyword that are exported.

Export modules
If creating tables for ECLIPSE BlackOil, the output from the Composition versus Depth
experiment can be passed through any separator. The default is to use a single stage at
standard conditions (that is, a stock-tank only).
Export Water
Description
This panel exports water properties.
The keyword PVTW or PVTWSALT can be exported for the ECLIPSE simulators. The
keyword
Hint All the keywords are described in the "ECLIPSE Reference Manual". WATPVT
is described in the "VFPi User Guide".
Units
Allows you to export in the ECLIPSE Unit set of choice.
Reservoir temperature
This sets the reservoir temperature to be used in generating the keywords.
Reservoir pressure
See above.
Dissolved natural gas

If this option is checked, the properties account for the presence of dissolved gas in the
water.
Brine
If this option is checked, salt concentration(s) can be accounted for.
VFPi
If this option is checked, the exported keyword will be WATPVT. The low temperature
and low pressure fields become active, these values correspond to the lower values to
be used in the flow table (for example, the top of the pipe).
For more information see "Water properties" on page 382.

Export modules
VFP module
Introduction
The VFP module generates blackoil tables for VFPi. To open this menu, select PVTi: File
| Export | VFPi from the main PVTi window.
Figure 6.7 The VFP module
The VFP module has the following options:

• File
This option allows you to read and display data from the PVI file. For further
information see "File" on page 150.
• Edit
This option allows you to add a new graph and copy items to the clipboard.
• View
This option allows you to control the appearance of the plot workspace.

VFP
• Define
This option allows you to define experiments and the method of calculation. For
further information see "Define" on page 151.
• Generate
This option allows you to generate tables. For further information see "Generate"
on page 153.
• Options
Allows the modification of the plot workspace.
• Help
Gives you access to the on-line help for this module.
VFP toolbar
The module toolbar contains the following buttons:
• Display section from PVI file

Use this option to display VFP sections from the PVI file. For further information
see page 150.
• Read section from PVI file
Use this option to load VFP sections from the PVI file. For further information see
page 151.
• Experiments
Use this option to define VFP experiments. For further information see page 152.
• Generate
Use this option to select the method of generation for the VFPi blackoil tables. For
further information see "Method" on page 152.
• Simulate
Use this option to generate the tables. For further information see "Perform" on
page 153.
• View results
Use this option to review the tables.
• View plots
Use this option to view plots of the tables. For further information see page 153.
File
Use this menu to read and display VFP sections from a PVI file.
The File menu has the following options:
Display VFP section from PVI...

Use this option to display a VFP section from the current PVI file.
Displaying VFP sections

1 Select VFP | File | Display VFP section from PVI... .
2 Select the VFP section.

VFP
Read VFP section from PVI...
Use this option to read data from a PVI file. You can use this option to load all or part
of the data required for the VFP section. Additionally you can choose to echo (the
contents of the whole PVI file to the current printfile, PVP.
Loading PVI data

1 Choose VFP | File | Read VFP section from PVI... and select the appropriate PVI
file.
PVTi searches for the required file, and looks for all occurrences of the required
section. If it finds more than one section of the required type in the file, you must
select which section you wish to read. The program displays a prompt specifying
the number of sections found.
2 Select the section you wish to load.
Note If the program does not find a VFP section in the PVI file it produces an error
message and stops processing.
Print preview
Previews the printed plot in the main module area.
Print layout
Allows you to set the text styles and sizes for printing.
Print setup
Sets up the printer specifications. This option is specific to Windows.
Print
Prints the whole module print area, the main window and allows you to set the print
type, for example color, postscript.
Close
Use this option to close the VFP module. To close the VFP module, select VFP | File |
Close.
Note The data is not lost. If you reselect the module all data that was previously set
is available.
Define
This menu allows you to set up experiments and the method of calculation. To open
this menu, select VFP | Define.

VFP
The Define menu has the following options:
• Experiments...
Defines experiments. For further information see "Experiments..." on page 152.
• Method...
Defines the method of calculation. For further information see "Method" on page
152.
Experiments...
Use this option to define experiments.
Defining experiments
1 Select VFP | Define | Experiments... .
This opens the Constant Composition Experiment panel.
2 Define a Constant Composition Expansion (CCE) experiment.
This is a depletion experiment where the moles entering the wellbore must leave
the wellbore. For further information on defining experiments see "Data
requirements for the experiments" on page 129.
Hint The first temperature you give for the CCE should be the highest temperature
in the production string, which can safely be taken to be the reservoir
temperature. The second temperature you give should be the lowest
temperature in the production string, for example a sub-sea temperature of 4
°C .
Note You do not need to give a saturation pressure as the program gives this data.
It also calculates undersaturated properties for all pressures. See "Output for
ECLIPSE simulators" on page 372 for further details
3 Define the separator conditions that take the reservoir liquid and/or vapor to
surface conditions to define the various ratios in the blackoil tables.
Note Note that to comply with the blackoil definition strictly, the last stage in the
separator train should be at standard conditions.
4 Specify the composition of the injection gas using one of the currently defined
sample mnemonics.
5 Give the gas-oil ratios in which the lean gas is to be added to the reservoir fluid.
6 Press OK.
Method
Use this option to select the method of calculation. You can choose from Coats, or
Whitson and Torp method. The default is the Whitson and Torp method

VFP
Choosing the method of calculation
1 Select VFP | Define | Method... .
This opens the Switch Method of Generation panel.
2 Choose the appropriate method.
For further information on the Coats method, [Ref. 3], or the Whitson and Torp
method, [Ref. 6], see "Output for ECLIPSE simulators" on page 372.
Generate
This menu allows you to generate blackoil tables for VFPi. To open this menu, select
VFP | Generate.
The Generate menu has the following options:
• Perform
Generates the blackoil tables.
Perform
Use this option to generate blackoil tables.
Note You must define a depletion experiment and the separator configuration
(which only need be a one stage default system to standard conditions) to
generate the blackoil tables.
Generating blackoil tables

1 Select VFP | Generate | Perform.
This generates the blackoil tables OILPVT and WGASPVT. The program also writes
gas, oil and water densities, at standard conditions, to the VFPi SURFDENS
keyword.
Note If the process is successful the program also writes the data to the PVO file
which you can add to an VFPi input file.
Note Note each set of tables is repeated twice, for the set of pressure nodes defined
on the CCE experiment, at the specified high (reservoir) and low temperatures.
Water properties
Click Yes when asked if you wish to generate water properties, using the VFPi
keyword WATPVT.
This opens a a series of data entry panels where you can enter the data necessary for
the in-built correlations to generate the appropriate data.
Plot
Use this option to produce plots.

VFP
Producing plots
1 Select VFP | View | Plot.
You can generate any of the following plots:
• oil formation volume factor, B o
• gas formation volume factor B g
• oil viscosity μ o
• gas viscosity μ g
• oil R s
Help menu
Gives access to help with PVTi.
Help
Opens the context-sensitive on-line help panel.

VFP
Utilities
Introduction
This menu allows you to change or redefine various program settings. It has the
following options:
• Units...
This allows you to set unit types. For further information see "Units..." on page 155.
• Standard Conditions...
This allows you to set the standard temperature and pressure. For further
information see "Standard conditions..." on page 156.
• Program Option
This allows you to change the program options For further information see
"Program options" on page 156.
• Debug...
This allows you to set the debug flags. For further information see "Debug..." on
page 161.
Units...
This option allows you to select various unit types. You can:
• Choose the unit type within PVTi.
• Choose the temperature unit type; this can be different from the one selected by the
above.
• Set mole fractions or percentages.
• Select absolute or gauge pressures.
For further information on the units see "Units" on page 431.
Setting unit types

1 To define the units, select Utilities | Units. This opens the Set PVTi Unit Definitions
panel. Each unit type is selected by clicking on the corresponding radio button.
Choosing the unit type for PVTi

Select from Metric, Field, Lab and PVT-Metric.
Setting the temperature unit type

Select the temperature unit type from Kelvin, Celsius, Rankine and Fahrenheit.
Setting mole fractions or percentages

Select the required option.
Selecting absolute or gauge pressures

Select the required option.

Run
Standard conditions...
This option allows you to set the standard temperature and pressure.
Defining standard conditions

1 To set the standard conditions select Utilities | Standard Conditions....
This opens the Standard Conditions panel.
2 Enter the standard temperature.
3 Enter the standard pressure.
4 Click on OK.
Program options
This option allows you to set various program options. Various options which cannot
be set elsewhere in PVTi have been collected together under this option.
1 To set the program options, select Utilities | Program Option.
This opens the Set PVTi Program Options panel.
You can set the following:
Table 6.2 Set PVTi Program Options panel
Program Option Available choices

"Definition of Liquid Saturation in CCE" on • Sliq = Vliq/Vsat
page 157. • Sliq = Vliq/Vtot
See “Treatment of Volume Shifts” on • Independent variables
page 157. • Dependent
"Separator GOR Calculation" on page 158. • Separator Conditions
• Stock Tank Conditions
"Temperature dependence for volume shifts" • None
on page 158 • Linear expansion only
• Polynomial correlations
"Specify/Calculate density and molar volume • user units
units" on page 158 • gm/cc and cc/gm-mole
"Binary interaction coefficients for EoS" on • Katz-Firoozabadi
page 158 • Cheuh-Prausnitz
"Specific heat capacity coefficients and • No Output of Values
calorific values" on page 158 • Output Values to Screen/PVP
"Calculated compositions" on page 158 • No save to samples
• Allow Optional Save to Samples
"Component Library" on page 159 • Katz-Firoozabadi
• Old PVTi Library
"Experiment compositions" on page 159 • Output to Screen/PVP
• No Output to Screen/PVP
"Experiment results" on page 159 • Always Output to PVP
• Optionally Output to PVP
• Never Output to PVP
"Plot vectors" on page 159 • No Output to file
• Output to Graf .VEC files

Utilities
Table 6.2 Set PVTi Program Options panel (Continued)
Program Option Available choices

"Print file output" on page 159 • A4 format
• 132 characters wide
"Definition of GOR in Diff. Lib." on page 159 • Normal
• No Last Stage
• Incremental
• Volume of Oil at Pbub
"Definition of Oil relative volume in Diff. Lib." • Oil FVF = Voil(p)/Voil(stc)
on page 160 • Oil FVF = Voil(p)/Voil(pbub)
"Black oil table output" on page 160 • All Data
• Truncation at Saturation Pressure
"LBC viscosity coefficients" on page 160 • Keep Fixed
• Allow Change when Regressing
"Flash calculations" on page 160 • E300 Flash
• Old PVTi Flash
"Sample mole fractions when regressing" on • Keep Fixed
page 160 • Allow Change
"Phase Plot Algorithm" on page 160 • New Phase Plots
• Old Phase Plots
"Write Keywords for Batch Mode" on page 160 • No
• Yes
Definition of Liquid Saturation in CCE

The definition of liquid saturation in the Constant Composition Expansion experiment
varies from laboratory to laboratory. The standard definition is to quote liquid
saturation as volume of liquid at pressure p per volume of fluid at saturation volume
(usually referred to as the cell volume). However, some laboratories refer to S liq as the
volume of liquid at pressure p per the total volume of fluid at pressure p .
• Select the appropriate option from the available list.
Treatment of Volume Shifts

Available options are Dependent and Independent. Dependent means that you cannot
regress on the volume shifts. They are defined as a function of the other critical
properties for each component and so if critical properties change, for example during
a regression, then the volume shifts are dynamically altered to stay consistent with the
new component data.
If you use the Independent option then you are allowed to regress on the volume shifts
of the components. PVTi also de-couples the dependence of the volume shifts on the
critical parameters so that a change in Tc, Pc, etc. does not effect a volume shift value.
The default setting is the Dependent option.

Utilities
Separator GOR Calculation
Some laboratories choose to quote separator liquid volumes at stock tank conditions,
rather than the actual pressure and temperature of the separator stage at which the
liquid is produced.
The default setting is the volume quoted at the separator conditions.
Temperature dependence for volume shifts

The volume shift corrections applied to the three-parameter PR3 and SRK3 equations
of state assume that the mis-match in predicted and measured liquid density at some
reference conditions on a component-by-component basis can be used to correct
volumes at all other pressures and temperatures. In an attempt to account for the
known temperature dependence, two methods are available for modifying the volume
shifts.
You can modify the shifts by applying linear thermal expansion to all components,
with an attempt at correction for molecular weight, or calculated for light components
as a polynomial involving temperature, with heavy components being modified by
thermal expansion, but without a correction for molecular weight. The methods are
described in "Three-parameter equation of state" on page 335.
Specify/Calculate density and molar volume units

Overriding the current units convention, you may specify that liquid and gas densities
should be output in units of gm/cc and molar volumes in cc/gm-moles.
Binary interaction coefficients for EoS

As an alternative to the BICs of Katz and Firoozabadi, the correlation of Cheuh and
Prausnitz (see "Binary interaction coefficients" on page 353) can be used to calculate
hydrocarbon-hydrocarbon BICs. If selected, the pre-multiplying A - coefficient can be
used as a special regression variable, especially useful for matching saturation
pressures.
• Select the BIC coefficients you wish to use.
Specific heat capacity coefficients and calorific values

The calculation of ideal gas specific heat in ECLIPSE Compositional can be
accomplished by switching on this flag, which then outputs the coefficients used in the
temperature-dependent expansion on a component-by-component basis. Calorific
values of the components of the system are also output to ECLIPSE Compositional
using this option.
• Switch on the output of the coefficients and calorific values, if required.
PVTi does not produce the coefficients and calorific values, by default.
Calculated compositions
Compositions calculated during an EoS simulation of an experiment can be saved to
be used later as samples for further experiments, phase plots, etc. This might be useful
in swelling tests, separator calculations or estimating the variation of composition with
depth.

Utilities
• Select the save samples option, if required.
PVTi does not save these samples, by default.
Component Library
This option allows you to specify the component library to use. The choice is between
the Katz-Firoozabadi and the Oil PVTi Library. The Katz-Firoozibadi is the default and
recommended choice.
Experiment compositions
This option allows you to turn off lengthy output to the print file of liquid and vapor
compositions calculated in experiments.
• Switch off the output of the liquid and vapor compositions if they are not required.
PVTi prints the liquid and vapor compositions to the screen and PVP file, by
default.
Experiment results
On definition of fluid properties, completion of experiments, etc., you can optionally
choose to write all the data to the PVP print file.
• Select the appropriate output option.
By default, PVTi always writes all the data to the PVP print file; however, you can
choose to make this write optional or never done.
Plot vectors
For more advanced graphical manipulation of PVTi plots, you may choose to output
data vectors to a file in GRAF user vector format. The vectors are written when the
plots are performed.
1 Choose whether you wish to produce the vectors.
PVTi does not produce these vectors by default.
Print file output

Use this option to determine the format of the print (PVP) file.
1 Select the appropriate option from the available list.
PVTi uses A4 as the default paper size format for this file.
Definition of GOR in Diff. Lib.

This option allows you to alter the definition of the GOR calculation in a differential
liberation experiment. Three alternative definitions are available:
• Normal where the last stage to standard conditions is removed with the volume of
gas being normalized to the volume of oil at reservoir conditions
• The GOR is defined in increments, that is, at each stage of the depletion process.
• The GOR is the default GOR but normalized to the volume of oil at its bubble point
pressure rather than at STC.
See "Differential liberation" on page 356 for the precise definitions of these quantities.
PVTi uses option 1 by default.

Utilities
Definition of Oil relative volume in Diff. Lib.
This option allows you to define an alternative definition for the relative oil volume in
a differential liberation experiment. When turned on, the option normalizes the
volume of oil at each stage to the volume of initial oil at its bubble point rather than
standard pressure. Refer to "Differential liberation" on page 356 for a precise definition
of the alternative.
Black oil table output

Switching to the Truncation at Saturation Pressure option outputs and plots only the
black oil table data relating to pressure values at the saturation pressure and below.
If the default of All Data is used then for pressure values above the saturation pressure,
PVTi swells the fluid with vapor in order to raise PSAT to the required value.
PVTi outputs all data to the blackoil tables by default.
LBC viscosity coefficients

This option allows the coefficients used in the LBC viscosity correlation to vary when
regressing to any viscosity observations.
• Choose whether you wish to vary the coefficients.
PVTi fixes the coefficients by default.
Flash calculations
This option allows the Flash and Psat algorithms to be changed from the default
ECLIPSE Compositional algorithms to the old PVTi (pre-99B) algorithms.
• Choose to use the old PVTi algorithms for Flash and Psat.
Sample mole fractions when regressing

This option allows you to vary mole fractions of components in any sample when
using the special regression variables CHARMF and/or MIXING.
• This option must be set in the ON state for you to use these regression variables.
You must have characterized and/or user components, or be mixing samples, for
the variables to become of use.
Phase Plot Algorithm

This option allows the phase plot algorithm to be changed from the default New Phase
Plots to the pre-2000A Old Phase Plots algorithm.
• Choose to use the old pre-2000A phase plot algorithm.
Write Keywords for Batch Mode

This option allows a .PVI file to be prepared for use in batch by writing extra
keywords.
• Choose to instruct PVTi to write out the extra keywords required for use in batch
mode. By default this option is disabled.

Utilities
This option cannot be saved, which means every time you open a project it is disabled
by default. The reason for this is that when this option is enabled it writes many more
keywords than is necessary in interactive mode. You could easily forget that a project
had this option enabled when opening an existing project meaning that many more
keywords than necessary would continue to be written to the PVI file.
Debug...
This option sets the debug flags.
Note This is a programmer test facility to request additional information from the
program at a debug level. You do not need to set one or more of the debug flags
unless help is required in tracing an apparent anomaly.
Monitor option
This is a programmer test facility to trace the root of a problem in the program at a
subroutine level.
Note You do not need to need to set this flag ON unless asked to do so.
Window
This menu allows you to control the size and appearance of the program windows. To
select this menu, select Window from the main PVTi window.
Tile
This puts all the visible windows into a “tiled” formation.
Cascade
This puts all the visible windows into a “cascade” formation.
Minimize children
This option minimizes all windows except the main PVTi window.
Restore children
This opens all minimized windows so that they are visible.
Help
This menu gives you access to the help for PVTi. To select this menu, select Help from
the main PVTi window.
Help
This opens the on-line help panels.

Window
ToolTips enabled
The ToolTips provide a single line of text about each toolbar button when the mouse
pointer is stationary over the icon. This option turns the ToolTips feature on and off.
About PVTi -...

This provides brief information on the program code version.
Right mouse button menu

Clicking the right mouse button on one of the buttons in the project tree displays a
popup menu that provides short-cuts to some of the common operations used in PVTi.

Right mouse button menu
Batch system and keywords
This section focuses on the batch mode functionality of PVTi. First of all a general
overview is given of the batch mode. The next section explains the new functionality
for PVTi 2004A, which enables you to set up a batch mode file in interactive mode. The
third section list all of the different keywords supported within the batch mode. The 3
sections are:
• "General information" on page 163.
• "Preparing Batch Mode Files in Interactive Mode" on page 164.
• "Overview of all supported keywords in Batch Mode" on page 166.
General information
For the 2004A version of PVTi the batch mode has undergone a significant revamp.
Over the last few years the user interface of PVTi has evolved rapidly and many of the
PVTi sections are now no longer written out to the PVI file as they are no longer
required when PVTi reads in a file. For example a BLACKOIL section is no longer
written out by PVTi when black oil tables are exported using the user interface.
Although they can be inserted into the file manually it was felt that a more user
friendly way of constructing the .PVI files containing all the appropriate sections
required for the batch mode was needed.
The next section outlines how it is now possible to perform a workflow interactively in
order for PVTi to be able to reproduce this workflow in batch mode at a later time.
The way in which you run a file in batch mode has also changed. Pre-2004A you had
to enter the keyword TESTCASE anywhere in the RUNSPEC section of the PVI file. You
then launched batch mode the command line, for example using $pvti filename (if using
a PC).
This has changed for the 2004A release. It was felt that opening files and adding
keywords and then remembering to remove them at a later data was cumbersome. The
new way to run in batch mode is to launch PVTi from the command line and specify
the word ‘-batch’ before the filename in the line command instruction.
Hint To launch the file TEST.PVI to run in PVTi’s batch mode on a PC use the
command ‘$pvti -batch TEST.PVI’. On a UNIX machine ‘@pvti -
batch TEST.PVI’ runs the file TEST.PVI in batch mode.
Sometimes users have more than 1 version of PVTi installed. It is also possible to
specify which version of PVTi to use on the command line using the -ver command.
Hint On a PC, to specify the 2004A version of PVTi to run the file TEST.PVI in
batch mode, use the command ‘$pvti -ver 2004a -batch TEST.PVI’.

Preparing Batch Mode Files in Interactive Mode
Overview
As explained in the previous section this new functionality introduced for PVTi 2004A
enables you to generate files suitable for batch mode using PVTi’s interactive mode.
The idea is that you would go through a pre-defined workflow in interactive mode and
then subsequently be able to make PVTi automatically perform this workflow in the
batch mode.
Note The workflow must be pre-defined. The batch mode is not designed to be able
to reproduce workflows where you have been experimenting with a particular
project.
1 Start PVTi in interactive mode with the file that you eventually wish to run in batch
mode.
2 Save the file concisely by doing File | Save (concise)....
This erases the history of any workflows stored in the file.
3 Open the Options panel using the File | Utilities | Program | Options....
4 In PVTi 2004A a new option has been added at the bottom called Write Keywords
for Batch Mode. Select Yes and then close the panel.
5 Perform the required workflow and, when finished, save the file (but not
concisely) using the File | Save... option.
Note The file must not be saved concisely as this would erase all the history in the
.PVI file that PVTi uses to reproduce your workflow in batch mode.
6 The final task is to actually run the file in batch mode. To do this on a PC launch
PVTi from the command line using the statement ‘$pvti -batch filename’
where filename is the name of your PVTi project, for example TEST.PVI.
Note The word ‘-batch’ can be put after or before the filename but it has to be
somewhere on the command line in order to tell PVTi to run in batch mode.
When you are running in batch mode the program automatically sends printed output
such as experiment simulation results to a print file with the same root name as the
input .PVI file. For example, if the input file is CRUDE.PVI, the print file is
CRUDE.PVP. See "PVI file" on page 166 for further information. In batch mode all the
experiments are automatically simulated in the project by default. If any regression is
to be performed in the batch mode then the experiment simulation is performed after
this has been done.
If keywords are exported for ECLIPSE during the batch run then .PVO files are created
as normal, but they are named using a convention that is outlined in the next section.
A file is also created called BATCH_OUT.PVI. This is a saved version of your project
after all the steps in the workflow have been performed. It can be useful to have this
file after the batch run has finished if the fluid model has changed (for example during
regression) in your project during the course of the workflow

Constraints on the workflow
In this new interactive approach to creating batch mode files there are constraints on
the workflows that you can perform. It is envisaged that some of these constraints will
be removed for PVTi 2005A. However, despite these constraints, all of the commonly
used functionality within PVTi is supported.
Regression
You are allowed a maximum of two REGRESS sections in the batch mode file. This is so
that regression on both special and normal variables is possible. Any further regression
sections are ignored by the batch mode.
Splitting/Grouping
You are allowed two GROUP sections and two SPLIT sections in the file. If there are
REGRESS sections then one SPLIT and one GROUP section is allowed before the first
REGRESS section and one SPLIT and one GROUP section is allowed after this REGRESS
section.
Export
It is assumed that all exports would be performed at the end of the workflow. An
unlimited amount of export sections (such as BLACKOIL, OUTECL3 sections) are
allowed as long as they are after the last REGRESS section. Because multiple exporting
is allowed a naming convention has been invented to stop PVTi just writing each
exported .PVO file over the top off the last one that was written out.
The naming convention depends on what kind of export is being performed:
BLACKOIL
If a BLACKOIL section has been read in the .PVI file then the naming convention of the
.PVO file is
“filename_samplename_experimenttype_keyword1keyword2.PVO”
where:
filename is the rootname of the project,
samplename is the sample name used for the export,
experimenttype is the type of depletion experiment used in the export,
keyword1 is the name of the first keyword exported
and keyword2 is the name of the second keyword exported.
For example if the file CRUDE.PVI was used to export sample ZI based on experiment
DL1 using Live Oil and Dry Gas keywords then the name of the produced file would
be CRUDE_ZI_DLLIVEDRY.PVO.
OUTECL3
If an OUTECL3 section has been read in the .PVI file then the naming convention of the
outputted .PVO file is filename_samplename_FLUIDMODEL.PVO.
APITRACK
If an APITRACK section has been read in the .PVI file then the naming convention of
the outputted .PVO file is filename_experimenttype_APITRACK.PVO.

Overview of all supported keywords in Batch
Mode
Listed below is a summary of all the keywords supported in a .PVI file using the batch
mode:
Note The COMB, PSEUCOMP and VFP sections-type keywords cannot currently
be prepared in a .PVI file for batch mode using interactive mode.
• "PVI file" on page 166.

• "Keywords introducing sections" on page 167.
• "RUNSPEC section keywords" on page 167.
• "SYSTEM section keywords" on page 168.
• "SPLIT section keywords" on page 169.
• "GROUP section keywords" on page 169.
• "COMB section keywords" on page 170.
• "SIMULATE section keywords" on page 170.
• "REGRESS section keywords" on page 170.
• "BLACKOIL section keywords" on page 171.
• "PSEUCOMP section keywords" on page 171.
• "OUTECL3 section keywords" on page 172.
• "VFP section keywords" on page 172.
PVI file
The PVI file consists of a number of sections, each introduced by a section keyword.
• The first section must be RUNSPEC, which specifies the number of components, the
equation of state option and the run title.
• The SYSTEM section must follow the RUNSPEC section.
• Other sections may be in arbitrary order, and may occur more than once.
Using the PVI file

After each operation that redefines the system, which may be by splitting, grouping or
regression, the program rewrites the audit trail. When you exit from a session you can
choose to write the audit trail to a PVI file. This creates new sections with the names
MODSPEC and MODSYS (corresponding to RUNSPEC and SYSTEM). The program treats
these new sections in the same manner as the original definitions. When you load a
PVI file created in a previous PVTi session the program automatically searches for the
RUNSPEC and any subsequent MODSPEC sections.
The PVTi keywords are described in detail in Chapter 7.

Keywords introducing sections
Each of the sections in a PVI file has a specific keyword to introduce it into the file.
Table 6.3 details these keywords.
Table 6.3 Keywords for introducing sections
Keyword Comments Details

RUNSPEC This must be the first section. See page 262
SYSTEM This must follow the RUNSPEC section. See page 276
SPLIT See page 273
GROUP See page 216
COMB See page 189
SIMULATE See page 271
REGRESS See page 259
BLACKOIL See page 184
VFP See page 292
PSEUCOMP See page 257
OUTECL3 See page 250
Keywords by section
Each of the main sections, for example the RUNSPEC section, has its own specific
keywords. Note the following:
• The keywords UNITS, DEGREES and STCOND are normally be specified in the
SYSTEM section, but can occur elsewhere, and can occur more than once.
• The keywords DEBUG and OPTIONS can be specified in the RUNSPEC section or in
the SYSTEM, COMB, SIMULATE or REGRESS sections.
• The keyword MESSAGE can be specified anywhere, its function being merely to
echo the argument to the print file at the time of a batch run.
RUNSPEC section keywords

Table 6.4 details the keywords specific to the RUNSPEC section.
Table 6.4 RUNSPEC keywords

EOS Selects the equation of state. See page 203
LBC Selects Lohrenz-Bray-Clark viscosities. See page 224
NCOMPS Sets the number of components. See page 240
NEWPVI Outputs a new PVI file at the end of the batch run. See page 241
PEDERSEN Selects the Pedersen viscosity correlation. See page 255
PRCORR Selects the modified Peng-Robinson equation. See page 256
TITLE Sets the run title. See page 281
ECHO Includes the PVI file in the PVP file. See page 202

Keywords by section
Table 6.4 RUNSPEC keywords (Continued)

NOECHO Does not include the PVI file in the PVP file. See page 243
DEBUG Sets the Debug flags. See page 196
DEBUE Sets Debug flags. See page 195
OPTIONS Sets Options flags. See page 248
VERSION Indicates the version of PVTi See page 291
SYSTEM section keywords

"The fluid model" on page 109 explains how to set up the SYSTEM keywords using
menu options. Table 6.5 details the keywords specific to the SYSTEM section.
Table 6.5 SYSTEM keywords

DEGREES Specifies the temperature convention. See page 198
STCOND Specifies the standard conditions. See page 275
UNITS Specifies the unit convention. See page 284
CNAMES Sets the component names. See page 187
LNAMES Sets the library component names. See page 227
CHARACT Specifies the characterisation of the components. See page 186
SCT Defines the Semi-Continuous-Thermodynamic split. See page 269
ACF Defines the acentric factors. See page 179
ACHEUH Defines the Cheuh-Prausnitz A -coefficient. See page 180
BIC Defines the binary interaction coefficients. See page 182

CALVAL Defines the calorific values. See page 185
DEFBIC Defines the default binary interaction coefficients. See page 197
DREF Defines the reference densities. See page 200
LBCCOEF Specify the viscosity coefficients. See page 225,
LBC page 224
MW Specifies the molecular weights. See page 238
OMEGAA Specifies non-default Ω a values (optional). See page 247
OMEGAB Specifies non-default Ω b values (optional). See page 247
PARACHOR Specifies parachors. See page 252

PCRIT Defines the critical pressures. See page 253
SPECHA,B,C,D Defines the specific heat capacity coefficients. See page 272
SSHIFT Specifies dimensionless volume shifts for PR3 EoS. See page 274
TBOIL Specifies boiling points (for ZJ equation). See page 278
TCRIT Defines the critical temperatures. See page 279
THERMX Specifies thermal expansion coefficient for volume shifts. See page 280
TREF Defines the reference temperatures. See page 283

Keywords by section
Table 6.5 SYSTEM keywords (Continued)

VCRIT Specifies the critical volumes. See page 289
VCRITVIS Defines critical volumes used in LBC viscosity correlation. See page 290
ZCRIT Specifies critical Z -factors. See page 301
ZCRITVIS Defines critical Z -factors used in LBC viscosity See page 302
correlation
ZI Defines sample composition (primary sample). See page 303
SAMPLES Defines other samples, lean gas for swelling test. See page 265
SAMTITLE Defines long titles for other samples. See page 267
HYDRO Specifies hydrocarbon/Non-hydrocarbon. See page 221
SALINITY Defines the salinity of specified samples. See page 263
MIX Allows mixing of two samples to form a new sample. See page 234
SPLIT section keywords

Table 6.6 details the keywords specific to the SPLIT section.
Table 6.6 SPLIT keywords

CORRACF Selects the acentric factor correlation. See page 191
CORRCP Selects the critical property correlation. See page 192
FRAC Sets the number and distribution of new plus fractions. See page 212
MDP Sets the molar distribution parameters. See page 230
MWS Specifies the plus fraction molecular weight. See page 239
SG Specifies the plus fraction specific gravity. See page 270
SCT Defines the Semi-Continuous-Thermodynamic split. See page 269
WHIT Defines a Whitson splitting. See page 298
GROUP section keywords

Table 6.7 details the keywords for the GROUP section.
Table 6.7 GROUP keywords

COMBINE Specifies fractions to be grouped together. See page 190
GRPBYSAM Specifies which sample to group to. See page 219
GRPBYWGT Specifies grouping by molecular weight. See page 220
GRPBYMIX Specifies grouping by mixing rule. See page 218
GRPBYALL Specifies grouping to average of all samples. See page 217

Keywords by section
COMB section keywords
The "COMB - Compositional Material Balance" on page 123 allows you to set up the
COMB keywords using menu options. Table 6.8 details the keywords for the COMB
section.
Table 6.8 COMB keywords

EXP Defines experiments. See page 205
OBS Defines observations. See page 244
OBSIND Defines the weight for individual observations. See page 245
COATS Requests Coats’ method for blackoil tables. See page 188
PEARCE Requests Pearce’s method for blackoil tables. See page 254
WHITSON Requests Whitson’s method for blackoil tables. See page 299
RECOVERY Sets the liquid production relative permeability for recoveries. See page 258
SIMULATE section keywords

"Simulation using PVTi" on page 128 describes how to set up the SIMULATE keywords
using menu options. Table 6.9 details the keywords for the SIMULATE section.
Table 6.9 SIMULATE keywords

FVFREF Sets the reference values for FVF calculations. See page 214
SAVCOMP Saves the information for calculated compositions. See page 268
REGRESS section keywords

"Regression in PVTi" on page 137 describes how to set up the REGRESS keywords using
menu options. Table 6.10 details the keywords for the REGRESS section.
Table 6.10 REGRESS keywords

VAR Defines variables. See page 285
FVFREF Sets the reference values for FVF calculations. See page 214
FIT Perform regression. See page 211
MAXIT Specifies the maximum number of iterations. See page 228

Keywords by section
Table 6.10 REGRESS keywords (Continued)

MAXSTEP Specifies the maximum regression step limit. See page 229
MINSTEP Specifies the minimum regression step limit. See page 233
REGTARG Specifies the regression target. See page 260
BLACKOIL section keywords

This section allows you to generate output for ECLIPSE Black Oil. The "Exporting
keywords" on page 144 describes how to set up the BLACKOIL keywords using menu
options. Table 6.11 details the keywords for the BLACKOIL section.
Table 6.11 BLACKOIL keywords

UNITS Defines the units to use in the export. See page 284
Note you can only define CVD, DL and SEPS.
WHITSON Requests Whitson and Torp’s method for blackoil tables. See page 299
DIFFERENTIAL Requests Differential method for blackoil tables. See page 199
FRAGOR Requests Fragor method for blackoil tables. See page 213
MOSES Requests Moses method for blackoil tables. See page 237
LIVEOIL Generates tables with live oil. See page 226
DEADOIL Generates tables with dead oil. See page 194
WETGAS Generates tables with wet gas. See page 297
DEADGAS Generates tables with dead gas. See page 193
WAT100 Outputs water properties. See page 293
MINDELP Specifies minimum compressibility test pressure See page 232
difference.
PSEUCOMP section keywords

The "Exporting keywords" on page 144 describes how to set up the PSEUCOMP
keywords using menu options. This section allows you to generate output for ECLIPSE
Black Oil options. Table 6.12 details the keywords for the PSEUCOMP section.
Table 6.12 PSEUCOMP keywords

Note you can only define CVD and SEPS.
GI Defines GI nodes for the GI option tables. See page 215

Keywords by section
Table 6.12 PSEUCOMP keywords (Continued)

MINELP Specifies minimum compressibility test pressure difference. See page 232
OUTECL3 section keywords

This section allows you to generate output for ECLIPSE Compositional. The
"Exporting keywords" on page 144 describes how to set up the OUTECL3 keywords
using menu options. Table 6.13 details the keywords for the OUTECL3 section.
Table 6.13 OUTECL3 keywords

UNITS Specifies the units to use in the export. See page 284
SAMPLE Specifies the fluid sample to use in the export. See page 264
RTEMP Defines the reservoir temperature. See page 261
NEWPVO Defines a rootname for the export file. See page 242
EOSOUT Requests equation of state data for ECLIPSE Compositional. See page 204
Note you can only define COMPG or CVD.
KVTABLE Requests K -value table for ECLIPSE Compositional. See page 222
XMFVP Requests XMFVP and YMFVP tables for ECLIPSE See page 300
Compositional.
YMFVP Requests XMFVP and YMFVP tables for ECLIPSE See page 300
Compositional.
ZMFVD Requests ZMFVD table for ECLIPSE Compositional. See page 304
VFP section keywords

This section allows you to generate output for VFPi. The "VFP module" on page 149
describes how to set up the VFP keywords using menu options. Table 6.14 details the
keywords for the VFP section.
Table 6.14 VFP keywords
Keywords Comments Details

Note you can only define CCE and SEPS.

Keywords by section
Table 6.14 VFP keywords (Continued)
Keywords Comments Details

WATVFP Outputs water properties. See page 296
TLOW Defines lowest temperature in the production string. See page 282
APITRACK section keywords

This section allows you to generate a series of black oil tables suitable for use with the
API Tracking option in ECLIPSE BlackOil. The "Export for API Tracking option in
ECLIPSE BlackOil" on page 145 describes how to set up the APITRACK keywords in
interactive mode using the appropriate panel. details the keywords for the APITRACK
section...
Table 6.15 APITRACK keywords

UNITS Defines the units to use in the export. See page 284
Note you can only define CVD, DL and SEPS.
SAMPLES The list of samples to export tables for. See page 266
DIFFERENTIAL Requests Differential method for blackoil tables. See page 199
FRAGOR Requests Fragor method for blackoil tables. See page 213
MOSES Requests Moses method for blackoil tables. See page 237
LIVEOIL Generates table with live oil. See page 226
DEADOIL Generates tables with dead oil. See page 194
WETGAS Generates tables with wet gas. See page 297
DRYGAS Generates tables with dry gas. See page 193
ALLDRY Generates gas tables for each sample. See page 181

Keywords by section
Error handling
Keyword errors
If an error occurs during keyword input, the program displays the offending line with
? characters under the field that is causing difficulties.
The error numbers given by the program may be of two types:
• Up to and including 100
These are errors detected within the PVTi data parser. Descriptions of these errors
are given in Table 6.16.
• Above 100
System data errors, usually caused by internal read operations used to convert a
character to its value.
Table 6.16 Error codes
Error code Description

1 Unable to read next line from data file.
2 Length of current line is zero.
3 Quote misplaces in stack or data field.
4 / character found in the wrong place.
5 Data field length greater than 24.
6 End of file reached.
7 Invalid character found.
8 Zero length stack found.
9 Stack length 1 with illegal character.
10 Incorrect stack pointers.
11 Failed to find * character in repeated data field.
12 Multiplier for repeated data field has more than 24 characters.
13 Error after internal read of multiple value.
14 Negative multiplier for repeated data field.
15 * character found as first token in stack.
16 Unrecognisable character for number in repeated data field.
17 Single field has the wrong data type.
18 Empty token next on stack.
19 Unknown data type.
20 Data read which is not of the type expected and which cannot be easily
converted.
21 Error in internal read to convert character data to number.
22 Repeated data has field length of more than 24 characters.
23 Error in internal read of repeated data multiplier.
24 More than one exponential character found.

Error handling
Table 6.16 Error codes (Continued)
Error code Description

29 Previous data type of repeated field not compatible for current repeated
data field.
30 Error in internal read of previous repeated data field.
31 Unable to convert current value due to error in previous repeated data field.
32 Values for margin setting not valid.
33 Zero length string found internally.

Error handling
Error handling
Keywords
Chapter 7
PVTi keywords
This chapter contains details of all the keywords in PVTi. The keywords are listed in
alphabetical order.
• "Keywords A-D" on page 178
• "Keywords E-K" on page 201
• "Keywords L- O" on page 223
• "Keywords P- S" on page 251
• "Keywords T - Z" on page 277.
Keywords 177
PVTi keywords
Keywords A-D
This section contains the A-D keywords.
• "ACF Acentric factors" on page 179
• "ACHEUH A-coefficient for Cheuh-Prausnitz BICs" on page 180
• "ALLDRY Dry Gas Tables for Each Sample" on page 181
• "BIC Binary interaction coefficients" on page 182
• "BLACKOIL Start of the BLACKOIL section" on page 184
• "CALVAL Specify calorific values" on page 185
• "CHARACT Components to be characterized" on page 186
• "CNAMES Component names" on page 187
• "COATS Blackoil tables" on page 188
• "COMB Start of the COMB section" on page 189
• "COMBINE Group existing components" on page 190
• "CORRACF Splitting correlation for ACFs" on page 191
• "CORRCP Splitting correlation for critical properties" on page 192
• "DEADOIL Dead oil tables" on page 194
• "DEBUE Select output to debug file" on page 195
• "DEBUG Select output to debug file" on page 196
• "DEFBIC Default binary interaction coefficients" on page 197
• "DEGREES Temperature convention" on page 198
• "DIFFERENTIAL Blackoil tables" on page 199
• "DREF Reference densities" on page 200.
• "DRYGAS Dry gas tables" on page 193
The other PVTi keywords are listed as follows:

"Keywords E-K" on page 201
"Keywords L- O" on page 223
"Keywords P- S" on page 251
"Keywords T - Z" on page 277.
178 Keywords
Keywords A-D
ACF Acentric factors
Specifies the acentric factors for the components in the system. The keyword is
RUNSPEC
X SYSTEM followed by N c values, and terminated with a slash character (/), where N c is the
SPLIT number of components specified in the RUNSPEC section.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
Example
OUTECL3
VFP ACF
.22500 .40000E-01 .13000E-01 .98600E-01 .15240 .18480
.20100 .22230 .25390 .30070 .27420 .30560
.34540 .38610 .42510 .46220 .49840 .53420
.57400 .61270 .64860 .68550 .72460 .76340
.80160 .84160 .88300 .92410 1.0590
/
Keywords 179
ACF
ACHEUH A-coefficient for Cheuh-Prausnitz BICs
Specifies the value of the A -coefficient (pre-multiplying) of the Cheuh-Prausnitz
RUNSPEC
X SYSTEM binary interaction coefficients for hydrocarbon-hydrocarbon interactions. Its value
SPLIT generally lies between 0.1 and 0.5.
GROUP
COMB
SIMULATE Note Only available if OPTIONS flag 6 is active.
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
--Default value:ACHEUH
0.15 /
180 Keywords
ACHEUH
ALLDRY Dry Gas Tables for Each Sample
This keyword requests that ECLIPSE BlackOil dry gas tables be output for each
RUNSPEC
SYSTEM sample. If this keyword is not present then a dry gas table is written out for the sample
SPLIT with the median vapor density at surface conditions.
GROUP
COMB The keyword ALLDRY has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
Keywords 181
DEADOIL
BIC Binary interaction coefficients
Specifies binary interaction coefficients. There are two possible formats for this
RUNSPEC
X SYSTEM keyword.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
First format
BLACKOIL
PSEUCOMP • The keyword should be followed by up to (N c -1) lines of data followed by a
OUTECL3
VFP forward slash character (/).
• Each line should consist of three integers followed by the set of binary interaction
coefficients between two specified components, terminated with a forward slash
character (/). The first integer specifies the index of the first component and the
second and third integers the lower and upper indices of the other component.
Second format
The program uses this format when saving.
This format in effect displays the lower triangle of the BIC matrix, and there is no need
for the first three columns informing the program of the component and the lower and
upper indices. PVTi assumes the coefficients are read in as second, third component,
etc., by row with the columns as the first, etc., component.
Note For both formats the interaction coefficient between any given component and
itself must be zero. PVTi always ensures that this is the case.
Examples
First format
Binary interaction coefficients for a six component system.
BIC
2 1 1 -0.02000 /
3 1 2 0.10000 0.03600 /
4 1 2 0.13000 0.05000 /
5 1 2 0.13500 0.08000 /
6 1 5 0.12700 2* 2*0.0600 /
/
Here there are zero binary interaction coefficients between components 4 to 3, 5 to 3,4
and 6 to 2,3.
182 Keywords
BIC
Second format
The same system of BICs appears as follows:
BIC
-0.02000
0.10000 0.03600
0.13000 0.05000
0.13500 0.08000
0.12700 2* 2*0.0600
/
Keywords 183
BIC
BLACKOIL Start of the BLACKOIL section
This is a delimiter keyword, specifying the start of the BLACKOIL section.
RUNSPEC
SYSTEM
SPLIT Note If present, this section must not appear before the SYSTEM section.
GROUP
COMB
SIMULATE
REGRESS
This section is used to generate blackoil tables for any of the currently defined fluid
X BLACKOIL samples by simulating reservoir depletion with a constant volume depletion or
PSEUCOMP differential liberation. Tables are generated that can then be input into ECLIPSE
OUTECL3
VFP
BlackOil.
184 Keywords
BLACKOIL
CALVAL Specify calorific values
Specifies the calorific values for the components in the system. The keyword is
RUNSPEC
X SYSTEM followed by N c values, where N c is the number of components specified in the
SPLIT RUNSPEC section. It is terminated with a / character. The current units convention is
GROUP
COMB used.
SIMULATE
-ml
REGRESS
BLACKOIL
The default is kJ ⁄ kg kJ/kg-ml. For alternatives a UNITS keyword should have been
PSEUCOMP previously read. For further information see [Ref. 39].
OUTECL3
VFP
Example
For a two component C1/I-C4 system, using PVT-Metric units:
-- Calorific Values kJ/kgmol

CALVAL
8.130000000E+02 2.659000000E+03 /
Keywords 185
CALVAL
CHARACT Components to be characterized
Specifies the components that are to be defined by characterization. This keyword is
RUNSPEC
X SYSTEM followed by up to N c lines (or as many non-blank components as in the CNAMES
SPLIT keyword, below), each terminated with a / character. The last line of data is followed
GROUP
COMB by a / character.
SIMULATE
REGRESS Each line begins with the mnemonic of the component to be characterized, followed by
BLACKOIL five items of data:
PSEUCOMP
OUTECL3 • Molecular weight M w
VFP
• Specific gravity S g
• Normal boiling point temperature T b
• Watson characterisation factor K w
• Reference temperature T ref
• Two strings which specify the characterisation procedure required for:

• Critical properties.
Kesler-Lee (K), Cavett (C), Riazi-Daubert (R), Winn (W) or Pedersen (P)
• Acentric factor.
Kesler-Lee (K), Edmister (E), Thompson (T) or Pedersen (P).
Generally, you must provide two out of the four data items M w ,S g ,T b and K w to
characterize a component. If more than two items are available, give the “best” two
items since the order of entry dictates which two are used. It is possible to perform a
characterization with a minimum data entry of M w ; the program estimates the specific
gravity from a Single Carbon Number (SCN) table look-up. See [Ref. 18].
Note If no reference temperature is specified it is given the value of the standard

temperature.
Example
Characterizes two plus fractions, the first of which has known specific gravity and
normal boiling point temperature, the second of which has only molecular weight.
Riazi-Daubert critical properties and Edmister acentric factor correlations to be used:
CHARACT
C7+1 1* 0.7500 390.0 2* R E /
C7+2 200.0 /
/
If more than two of the four items, M w ,S g ,T b and K w are available, then the program
uses the first two, that is given data for M w ,T b and K w , then PVTi uses M w and T b .
186 Keywords
CHARACT
CNAMES Component names
Specifies the mnemonics associated with the components in the system that are either
RUNSPEC
X SYSTEM to be characterized or fully user defined. This keyword is followed by N c component
SPLIT
GROUP
names, where N c is the number of components specified in the RUNSPEC section. It is
COMB terminated with a / character.
SIMULATE
REGRESS Define library components using the LNAMES keyword. Any system containing both
BLACKOIL
PSEUCOMP
library and characterisation/user defined components should contain both LNAMES
OUTECL3 and CNAMES with the default specification 1*, where appropriate.
VFP
The program translates the component names into upper case on input. The length
may be up to 72 characters, but a limit of four is suggested to fit into the program
output formats.
Example
For a nine-component condensate system, the first five being from library, the last four
to be user defined:
CNAMES
5* C4-6 C7+1 C7+2 C7+3 /
LNAMES
CO2 N2 C1 C2 C3 4* /
Keywords 187
CNAMES
COATS Blackoil tables
Specifies that the Coats method for the generation of blackoil tables is used in
RUNSPEC
SYSTEM preference to the Whitson and Torp method. By default, the program uses the Whitson
SPLIT and Torp method.
GROUP
X COMB The keyword COATS has no arguments.
SIMULATE
REGRESS
X BLACKOIL
X PSEUCOMP
OUTECL3
X VFP
188 Keywords
COATS
COMB Start of the COMB section
This is a delimiter keyword, specifying the start of the COMB section.
RUNSPEC
SYSTEM
GROUP
X COMB
SIMULATE
REGRESS
This section is used for data and consistency checking of laboratory data of gas
BLACKOIL condensates and volatile oils. Material balance is performed on any Constant Volume
PSEUCOMP Depletion data, to calculate liquid compositions, K-values, densities, etc. Additionally,
OUTECL3
VFP
tests of separator recombination data are available, as well as estimates of recovery, and
generation of blackoil tables without recourse to the equation of state.
Keywords 189
COMB
COMBINE Group existing components
Requests the grouping of a range of existing components into a new combined fraction.
RUNSPEC
SYSTEM This is done by specifying a group index for each existing component. PVTi groups
SPLIT together components with the same group index.
X GROUP
COMB Components not to be grouped may be ignored, or given a group index of zero.
SIMULATE
REGRESS
BLACKOIL Note The group index has no significance other than to identify sets of components.
PSEUCOMP
OUTECL3
VFP You may enter a name for the new fractions. By default the program uses GRP1, GRP2
and GRP3 as the group names.
The form of the keyword is:
COMBINE
icl1 icu1 ifr1 zfr1 /
icl2 icu2 ifr2 zfr2 /
/
where:
• icl1 is a lower existing component index;
• icu1 is an upper existing component index;
• ifr1 is a group index;
• zfr1 is the mnemonic for a fraction.
Example
In a 19-component system, group the components 10 to 14 inclusive, and the
components 15, 17 and 18. Accept default names for the fractions:
COMBINE
10 14 1 /
15 15 2 /
17 18 2 /
/
190 Keywords
COMBINE
CORRACF Splitting correlation for ACFs
Specifies the correlation to be used for calculating the acentric factors of characterized
RUNSPEC
SYSTEM split fractions. It should be either Kesler-Lee (K), Edmister (E), Thomassen (T) or
X SPLIT Pedersen (P) [Ref. 13], [Ref. 14], [Ref. 30] & [Ref. 42] respectively. It must be terminated
GROUP with a forward slash (/) character.
COMB
SIMULATE The default is Kesler-Lee (K).
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP Example
Edmister correlation required:
CORRACF
E /
Keywords 191
CORRACF
CORRCP Splitting correlation for critical properties
Specifies the correlation to be used for calculating the critical properties of
RUNSPEC
SYSTEM characterized split fractions. It should be one of Kesler-Lee (K), Cavett (C) Riazi-
X SPLIT Daubert (R), Winn (W), or Pedersen (P) [Ref. 10], [Ref. 11], [Ref. 12], [Ref. 43] , [Ref. 44],
GROUP and [Ref. 42] respectively. It should be terminated by a forward slash (/) character.
COMB
SIMULATE The default is Kesler-Lee (K).
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP Example
Cavett correlation required:
CORRCP
C /
192 Keywords
CORRCP
DRYGAS Dry gas tables
This keyword requests that ECLIPSE BlackOil dry gas tables be output.
RUNSPEC
SYSTEM The keyword DRYGAS has no arguments.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
Keywords 193
DEADGAS
DEADOIL Dead oil tables
This keyword requests that ECLIPSE BlackOil dead oil tables be output.
RUNSPEC
SYSTEM The keyword DEADOIL has no arguments.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
194 Keywords
DEADOIL
DEBUE Select output to debug file
Requests that additional output from the PVTLIB annex be directed to the DBG debug
X RUNSPEC
SYSTEM file. The keyword is followed by up to eight integers in the range one to eight,
SPLIT terminated with a forward slash (/) character. The indices request the following
GROUP output:
COMB
SIMULATE
REGRESS
BLACKOIL Table 7.1 Output indices
PSEUCOMP
OUTECL3 Index Output from
VFP
0 No debug
1 Full debug (see below)
2 Flash stability routines
3 Flash split routines
4 Newton routines
5 Fugacity routines (enter as -1 to obtain fugacities of components,
batch mode only)
6 Water EoS routines
7 p sat routines
8 Critical point routines

9 Phase plot generator
Example
Request debug from Newton and p sat routines:
CORRCP
C /
Keywords 195
DEBUE
DEBUG Select output to debug file
Requests that additional output from PVTi be directed to the DBG debug file. The
X RUNSPEC
SYSTEM keyword is followed by up to 16 integers in the range one to 16, terminated with a
SPLIT forward slash (/) character. The indices request the following output:
GROUP
COMB
SIMULATE
REGRESS Table 7.2 Output indices
BLACKOIL
PSEUCOMP Index Output from
OUTECL3
VFP 1 Array allocation routines
2 Phase plots
3 Zudkevitch-Joffe equation of state
4 Regression
5 Parser
6 Characterization/Splitting
7 Experiments (not p sat )
8 Observations
9 Saturation pressure/temperature
10 Flash routine
11 Newton routines
12 Fugacity evaluation (set as -1 to obtain Fugacities of components -
batch mode only)
13 Stability Check/Michelsen
14 Opening/closing of files
15 Regression output
16 LBC coefficients with regression
17 COMB routines
Example
Request debug from Flash and Michelsen routines only:
DEBUG
9* 10 2* 13 /
196 Keywords
DEBUG
DEFBIC Default binary interaction coefficients
Specifies that default binary interaction coefficients are required. These are only
RUNSPEC
X SYSTEM available between library/library and library/characterization components.
SPLIT
GROUP
The keyword DEFBIC has no arguments.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 197
DEFBIC
DEGREES Temperature convention
Specifies the temperature convention to use. The keyword is followed by a character
RUNSPEC
X SYSTEM string identifying the convention, which may be one of the following four:
SPLIT
GROUP
• KELVIN
COMB Degrees Kelvin
SIMULATE
REGRESS • CELSIUS
BLACKOIL Degrees Celsius
PSEUCOMP
OUTECL3 • RANKINE
VFP Degrees Rankine
• FAHRENHEIT
Degrees Fahrenheit.
Only the first character is significant, and may be lower or upper case. The default
temperature convention is degrees Kelvin.
Note This keyword must follow the UNITS keyword.
Example
Set to degrees Fahrenheit:
DEGREES
F /
198 Keywords
DEGREES
DIFFERENTIAL Blackoil tables
Specifies that differential black oil tables are generated. By default, the Whitson and
RUNSPEC
SYSTEM Torp method of composite black oil table generation is used.
SPLIT
GROUP
The keyword DIFFERENTIAL has no arguments.
COMB
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 199
DIFFERENTIAL
DREF Reference densities
Specifies the reference densities for the components in the system. The keyword is
RUNSPEC
X SYSTEM followed by N c values, and terminated with a forward slash (/) character, where N c is
SPLIT the number of components specified in the RUNSPEC section.
GROUP
COMB
SIMULATE
REGRESS Note These reference densities are only required if the Zudkevitch-Joffe or three-
BLACKOIL parameter Peng-Robinson equations have been selected (although it is good
PSEUCOMP practice to always enter them), and are associated with the reference
OUTECL3
VFP temperatures specified with the TREF keyword.
• UNITS: 3
METRIC: kg/m FIELD: lb/cu ft
3
LAB: gm/cc PVT-M: kg/m
Example
For a two-component CO2/I-C4 system:
DREF
777.00 557.00 /
200 Keywords
DREF
Keywords E-K
This section contains the E-K keywords.
• "ECHO Insert PVI file into PVP file" on page 202
• "EOS Defines the required Equation of State" on page 203
• "EOSOUT EoS data for ECLIPSE 300" on page 204
• "EXP Experiments" on page 205
• "EXPIND Set Status of Experiments" on page 210
• "FIT Perform fit by regression" on page 211
• "FRAC Specify plus fraction data" on page 212
• "FRAGOR Blackoil tables" on page 213
• "FVFREF FVF reference conditions" on page 214
• "GI Define GI nodes for E200 tables" on page 215
• "GROUP Start of the GROUP section" on page 216
• "GRBYALL Start of the GROUP section" on page 217
• "GRBYMIX Start of the GROUP section" on page 218
• "GRBYSAM Start of the GROUP section" on page 219
• "GRPBYWGT Grouping by molecular weight" on page 220
• "HYDRO Define component as hydrocarbon or non-hydrocarbon" on page 221
• "KVTABLE Request K-value table for ECLIPSE 300 output" on page 222

"Keywords A-D" on page 178
Keywords 201
Keywords E-K
ECHO Insert PVI file into PVP file
This keyword inserts the PVI input file into the PVP print file.
X RUNSPEC
SYSTEM The keyword ECHO has no arguments.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
202 Keywords
ECHO
EOS Defines the required Equation of State
Defines the required equation of state from the list:
X RUNSPEC
SYSTEM • PR
SPLIT
GROUP
Two-parameter Peng-Robinson.
COMB • PR3
SIMULATE
REGRESS Three-parameter Peng-Robinson.
BLACKOIL
PSEUCOMP • RK
OUTECL3 Redlich-Kwong.
VFP
• SRK
Soave-Redlich-Kwong.
• SRK3
Three-parameter Soave-Redlich-Kwong.
• SW
Schmidt-Wenzel.
• ZJ
Zudkevitch-Joffe.
The default is PR3.
Example
EOS
PR3 /
Keywords 203
EOS
EOSOUT EoS data for ECLIPSE 300
Requests that equation of state data be output to the PVO file for inclusion in ECLIPSE
RUNSPEC
SYSTEM Compositional. The EOSOUT keyword has no arguments.
SPLIT
GROUP
By default, if this keyword is omitted and no other keywords are in the OUTECL3
COMB section, EoS data is automatically output to the PVO file.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
X OUTECL3
VFP
204 Keywords
EOSOUT
EXP Experiments
Specifies the experiments to be performed. Up to 100 experiments plus one set of
RUNSPEC
SYSTEM BLACKOIL calculations can be defined.
SPLIT
GROUP
The form of the keyword and associated data is:
X COMB
X SIMULATE EXP
X REGRESS
<no>1 <samp>1 <type>1 <data>1 /
X BLACKOIL
X PSEUCOMP <no>2 <samp>2 <type>2 <data>2 /
X OUTECL3 :
X VFP
where:
• <no> specifies the experiment number, usually increasing sequentially from 1;
• <samp> specifies the mnemonic of the required sample composition for this
experiment;
• <type> specifies the experiment type. One of: FLASH, BUBBLE, DEW, CCE, CVD, DL,
SWELL, SEPS, COMPG, TERN, vapor, MFLASH, CRIT, TSAT, FCMP, MCMP, PSAT, MCT.
In the above, a forward slash (/) character terminates entry of experiments, and the
form of the <data> depends on the particular experiment type:
Table 7.3 Required data for experiments
Experiment Data
FLASH T P
BUBBLE T
DEW IRT T
CCE FLUID T P P2... ...PN
CVD FLUID T P1 P2... ...PN
DL T P P2... ...PN
SWELL FLUID YMF T [PRESS P] STYPE
X1... ...XN
SEPS T1 P1 L1 V1
(option 1) T2 P2 L2 V2
. . . .
TN PN LN VN
SEPS T1 P1 FL11 FV11 FV1N
(option 2) T2 P2 FL21 FV21...FL2N FV2N
. . . . .
TN PN LN FVN1...FLN FVNN
N
COMPG FLUID T P D D1
TERN T P ZG1 ZG2 ZG3 IG1
IG2... ...IGN
VAPOR FLUID YMF T P M1... ...MN
Keywords 205
EXP
Table 7.3 Required data for experiments (Continued)
Experiment Data
MFLASH T P
CRIT
TSAT P TSAT
FCMP YMF T FCMP
MCMP YMF ICORV T MCMP
PSAT T PSAT
MCT YMF ICORV T P FL2...FLN-1N MCT
(option 1) (option
1)
MCT YMF ICORV T P FV2...FVN-1N MCT
(option 2) (option
2)
TCLOUD T P
PCLOUD T P
where:
Table 7.4 Keyword arguments
Keyword
Data requirements
arguments
T Single temperature entry
P Single pressure entry
IRT Retrograde (IRT=1) or normal (IRT=0) dew point
FLUID GAS or OIL system
PI I’th pressure
YMF Mnemonic of gas sample composition defined in SAMPLES or ZI.
STYPE Specifies subsequent data is MOLE fraction of gas in the mixture or
GOR gas volume at STP in the oil at original pressure or other
pressure.
ICORV Condensing (ICORV=0) or vaporizing (ICORV=1) multiple contact
miscibility pressure required
XI Mole fraction or GOR (see STYPE) for step I.
MI Number of moles added cumulatively at step I.
TI Temperature of I’th separator stage.
LI Destination of liquid output from stage. Default is next stage, I+1.
VI Destination of vapor output from stage. Default is the cumulative
vapor output, 0.
Note, for separators, cumulative output stage is labelled 0, with
vapor at STC, liquid at final stage conditions.
FLIJ Fraction of liquid output from stage I to stage J. Default is 1.0 to the
next stage.
206 Keywords
EXP
Table 7.4 Keyword arguments (Continued)
Keyword
Data requirements
arguments
FVIJ Fraction of vapor output from stage I to stage J. Default is 1.0 to the
cumulative vapor output, 0.
D Reference depth.
DI Depth I.
ZGI Names of the I’th ternary group.
IGI Ternary group index (1-3) for component I.
Example
Consider the following experiments:
1 Dew point pressure at 200 °F , no pressure estimate given;
2 Swelling test at 200 °F , injection gas composition specified in SAMPLES with

mnemonic ZINJ, four stage addition with given mole percentage of gas in reservoir
fluid;
3 Constant composition expansion of condensate (GAS) at 200 °F , pressures in psi;
4 Constant volume depletion of condensate at 200 °F , pressures in psi;
5 2-stage separator at 500 psi, 125 °F and 50 psi, 90 °F with second stage fed from
liquid output of stage one;
6 Composition with depth: a volatile oil (OIL) at reference conditions of 220 °F , 4000
psi and 7500 ft. Require compositions and pressures at 7000, 7100,..., 8000 ft;
7 Ternary diagram calculation for an eight component system, explicit grouping;
8 Ternary diagram calculation for an eight component system, default grouping;
9 First contact miscibility pressure at 160 °F on ZI, injecting ZINJ;
10 Multiple contact miscibility pressure at 160 °F on ZI, injecting ZINJ, vaporizing

drive;
11 Multiple contact miscibility pressure at 160 °F on ZI, injecting ZINJ, condensing
drive;
12 Multiple contact test at 121.1 °F and 2738 psi on ZI, injecting ZINJ, condensing
drive. At each stage all of the remaining oil from the flash is contacted with one
mole of ZINJ.
Keywords 207
EXP
EXP
1 ZI DEW 1 200.0 /
2 ZI SWELL GAS ZINJ 200.0
MOLE 0.1271 0.3046 0.5384 0.6538 /
3 ZI CCE GAS 200.0 6000.0 5000.0 4000.0 3000.0
2000.0 1000.0 /
4 ZI CVD GAS 200.0 3000.0 2400.0 1800.0 1200.0 /
-- old format for SEPS
5 ZI SEPS 125.0 500.0 2 0
90.0 50.0 0 0 /
7 ZI TERN N2C1 C2C5 C6+ 1 1 2 2 2 2 3 3 /
8 ZI TERN /
9 ZI FCMP ZINJ 160.00 /
10 ZI MCMP ZINJ 1 160.00 /
11 ZI MCMP ZINJ 0 160.00 /
12 ZI MCT ZINJ 0 121.10 2738.0 1.0 1.0 1.0 1.0 /
/
Note At pressures above the saturation pressure where no compositional changes

take place, the Constant Volume Depletion (CVD) and the Differential
Liberation (DL) are effectively a Constant Composition Expansion (CCE)
experiment.
Restrictions
The EXP keyword arguments are subject to restrictions in the BLACKOIL, COMB,
OUTECL3, REGRESS and VFP sections:
Table 7.5 Restrictions for EXP keyword arguments
Section Experiment
BLACKOIL CVD, DL & SEPS only.
PSEUCOMP CVD & SEPS only.
COMB CVD, SEPS & CCE only.
OUTECL3 CVD & COMPG only.
REGRESS Not COMPG, TERN, MFLASH, CRIT, TSAT, FCMP, MCMP.
VFP CCE & SEPS only.
BLACKOIL section
In the BLACKOIL section, to generate blackoil tables you must define a depletion
experiment and a separator network.
Number these experiments; 1 (CVD or DL) and 2 (SEPS).
208 Keywords
EXP
COMB section
In the COMB section, you are restricted to the above experiments for the material
balance (CVD), recombination of separator data (SEPS) and recovery calculations
(CCE).
Number these experiments 1 to 3, in any order.
OUTECL3 section
In the OUTECL3 section, you are restricted to a CVD for the purposes of generating
either KVTABLE or XMFVP and YMFVP tabular data.
Note Only two pressures should be entered, the maximum and minimum pressures
required in the table. The other pressures are calculated by interpolation (up to
50 pressures in all).
You can also specify a COMPG experiment for the purpose of generating ZMFVD tables
for ECLIPSE Compositional.
Note Only two depths should be entered, the maximum and minimum depths
required in the table. The other depths are calculated by interpolation (up to 50
depths in all).
Number these experiments 1 (CVD) and 2 (COMPG).
VFP section
In the VFP section, to generate blackoil tables you must define a depletion experiment
and a separator network.
Number these experiments 1 (CCE) and 2 (SEPS).
Keywords 209
EXP
EXPIND Set Status of Experiments
This keyword is used to set the status of a particular experiment, necessary when
RUNSPEC
SYSTEM performing a regression operation. The keyword is a list of integers, one for each
SPLIT experiment, where a 0 means the experiment is not to be used in a regression (‘turned
GROUP off’) and a 1 means the experiment is to be used in a regression (‘turned on’).
X COMB
X SIMULATE
X REGRESS
X BLACKOIL
X PSEUCOMP Example
X OUTECL3
X VFP
We have 20 experiments in total in our project and wish that they all be ‘turned on’
during regression except the 5th, 10th, 14th, 16th and 18th, which will be ‘turned off’:
OPTIONS
1 1 1 1 0 1 1 1 1 0 1 1 1 0 1 0 1 0 1 1/
210 Keywords
OPTIONS
FIT Perform fit by regression
This keyword requests that a regression operation, previously specified using the EXP,
RUNSPEC
SYSTEM OBS and VAR keywords, be carried out.
SPLIT
GROUP
The keyword FIT has no arguments.
COMB
SIMULATE
X REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 211
FIT
FRAC Specify plus fraction data
Specifies the number and relative compositions of the plus fraction splitting. You
RUNSPEC
SYSTEM specify the number of fractions required, the relative composition of the fractions, and
X SPLIT the mnemonics to be associated with the split fractions (optional), terminated with a /
GROUP character.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP Example
OUTECL3
VFP
Split the plus fraction into three sub-fractions:
FRAC
3 0.72 0.23 0.05 C7+1 C7+2 C7+3 /
212 Keywords
FRAC
FRAGOR Blackoil tables
Specifies that the Fragor method for the generation of composite from differential
RUNSPEC
SYSTEM blackoil tables is used.
SPLIT
GROUP
By default, the program uses the Whitson and Torp method of composite black oil table
COMB generation.
SIMULATE
REGRESS The keyword FRAGOR has no arguments.
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 213
FRAGOR
FVFREF FVF reference conditions
The reference conditions for the liquid formation volume factors from Pinit/Pbub to
RUNSPEC
SYSTEM Pstock (mnemonic SRELV) calculation in separator experiments may be input here.
SPLIT
GROUP
For each SEP experiment, a reference temperature and pressure are required. If the
COMB reference pressure is zero, the saturation pressure at the reference temperature will be
X SIMULATE used.
X REGRESS
BLACKOIL Each record must be terminated with a / character. The set of records must end with a
PSEUCOMP blank record, containing only a slash (/).
OUTECL3
VFP
Note If reference conditions are not entered for a SEP experiment, no FVFs are
calculated for that experiment. This includes the FVFs at each stage, and the
ORELV mnemonic, as well as SRELV.
Example
FVFREF
2 220.0 4000.0 /
4 200.0 0.0 /
7 210.0 /
/
214 Keywords
FVFREF
GI Define GI nodes for E200 tables
Specifies the lean gas injection sample and the GI nodes (gas-oil ratios) for the
RUNSPEC
SYSTEM generation of ECLIPSE pseudo-compositional tables. This keyword is followed by a
SPLIT mnemonic specifying the fluid sample (defined on the SAMPLES keyword) and up to
GROUP ten GI values, terminated with a / character.
COMB
SIMULATE
REGRESS
BLACKOIL
X PSEUCOMP Example
OUTECL3
VFP
Inject sample ZINJ at GI’s of 0.1, 0.2, 0.3 mscf/rb:
GI
ZINJ 0.1 0.2 0.3 /
Keywords 215
GI
GROUP Start of the GROUP section
This is a delimiter keyword, specifying the start of the GROUP section.
RUNSPEC
SYSTEM
X SPLIT Note If present, this section must not appear before the SYSTEM section.
GROUP
COMB
SIMULATE
REGRESS
This section is used to define data necessary for the grouping of components into
BLACKOIL pseudo-components.
PSEUCOMP
OUTECL3
VFP
216 Keywords
GROUP
GRBYALL Start of the GROUP section
This keyword specifies that the grouping of components specified by the COMBINE
RUNSPEC
SYSTEM keyword be performed with respect to an average of all samples rather than to the
SPLIT default ZI sample.
X GROUP
COMB The keyword GRBYALL has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 217
GRBYALL
GRBYMIX Start of the GROUP section
RUNSPEC
SYSTEM keyword be performed using a molecular weighting, rather than the default mole
SPLIT fraction approach.
X GROUP
COMB The keyword GRBYMIX has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
218 Keywords
GRBYMIX
GRBYSAM Start of the GROUP section
RUNSPEC
SYSTEM keyword be performed with respect to the named sample, rather than to the default ZI
SPLIT sample.
X GROUP
COMB
SIMULATE
REGRESS
BLACKOIL Example
PSEUCOMP
OUTECL3 Group to sample ZINJ
VFP
GRPBYSAM
ZINJ /
Keywords 219
GRBYSAM
GRPBYWGT Grouping by molecular weight
RUNSPEC
SYSTEM keyword be performed using a molecular weighting, rather than the default mole
SPLIT fraction approach.
X GROUP
COMB The keyword GRBYWGT has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
220 Keywords
GRPBYWGT
HYDRO Define component as hydrocarbon or non-
hydrocarbon
Specifies whether a component is a hydrocarbon or a non-hydrocarbon. The keyword
RUNSPEC
X SYSTEM is followed by up to N c (the number of components in the fluid system) flags indicating
SPLIT the component types and terminated by a forward slash (/) character.
GROUP
COMB The available flags are given in Table 7.6.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP Table 7.6 Component Types
OUTECL3
VFP Type Flag in HYDRO
Non-hydrocarbon N
Hydrocarbon H
Paraffin P
Cyclic-hydrocarbon (naphthalene) C
Aromatic A
By default, all components are assumed to be hydrocarbons, except for specific non-
hydrocarbons in the “LIB” set, that is, CO 2 ,H 2 S ,H 2 O ,N 2 ,H 2 and CO .
Example
An 11 component system consisting of CO2, N2, C1, C2, C3, IC4, NC4, IC5, NC5, C6,
C7+:
HYDRO
N N 9*H /
Keywords 221
HYDRO
KVTABLE Request K-value table for ECLIPSE 300
output
Requests that a K -value table, consisting of the K -value of each component at a set of
RUNSPEC
SYSTEM pressures, be output to the PVO file. The keyword KVTABLE takes no arguments.
SPLIT
GROUP
COMB
SIMULATE
REGRESS Note Your must also define a Constant Volume Depletion (CVD) experiment with
BLACKOIL the EXP keyword and associated data in order that the KVTABLE can be
PSEUCOMP generated.
X OUTECL3
VFP
Note, also, that only two pressures need be defined in the CVD; these should
be the maximum and minimum required.
222 Keywords
KVTABLE
Keywords L- O
This section contains the L-O keywords.
• "LBC Lohrenz-Bray-Clark viscosities" on page 224
• "LBCCOEF Set non-default LBC coefficients" on page 225
• "LIVEOIL Live oil tables" on page 226
• "LNAMES Specify library names" on page 227
• "MAXIT Max. number of regression iterations" on page 228
• "MAXSTEP Maximum step size allowed in regression" on page 229
• "MDP Data for Whitson splitting" on page 230
• "MESSAGE Echo message to file and screen" on page 231
• "MINDELP Minimum pressure difference" on page 232
• "MINSTEP Minimum step limit allowed in regression" on page 233
• "MIX Mix samples" on page 234
• "MODSPEC Denotes start of the run specification section" on page 235
• "MODSYS Start of the MODSYS section" on page 236
• "MOSES Blackoil tables" on page 237
• "MW Specify molecular weights" on page 238
• "MWS Define plus fraction mole weight for CMF splitting" on page 239
• "NCOMPS Specify number of components" on page 240
• "NEWPVI Request new output PVI file" on page 241
• "NEWPVO Request new output PVO file" on page 242
• "NOECHO No insertion of PVI file into PVP file" on page 243
• "OBS Specify observations" on page 244
• "OBSIND Specify observation weights" on page 245
• "OMEGAA,OMEGAB Specify EoS omega values" on page 247
• "OPTIONS Set various program options" on page 248
• "OUTECL3 Start of the OUTECL3 section" on page 250

Keywords 223
Keywords L- O
LBC Lohrenz-Bray-Clark viscosities
Specifies that the Lohrenz-Bray-Clark viscosity correlation is to be used, rather than the
X RUNSPEC
SYSTEM Pedersen method. The LBC correlation is the default and is used if neither of the LBC
SPLIT or PEDERSEN keywords is used.
GROUP
COMB The keyword LBC has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
224 Keywords
LBC
LBCCOEF Set non-default LBC coefficients
Enables the default coefficients used by the Lohrenz-Bray-Clark viscosity correlation
RUNSPEC
X SYSTEM to be modified. This should only be done with great care: the viscosity is obtained from
SPLIT a fourth order polynomial in reduced density and must clearly not go negative.
GROUP
COMB There are five coefficients with default values of 0.1023, 0.023364, 0.058533, -0.040758
SIMULATE and 0.0093324. Any coefficient not specified with LBCCOEF takes these default values.
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP Example
Reset two of the LBC coefficients
LBCCOEF
1* 0.025 1* -0.04 0.01 /
Keywords 225
LBCCOEF
LIVEOIL Live oil tables
This keyword requests that ECLIPSE BlackOil live oil tables be output.
RUNSPEC
SYSTEM The keyword LIVEOIL has no arguments.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
226 Keywords
DEADOIL
LNAMES Specify library names
Specifies the mnemonics associated with the components in the system that are to be
RUNSPEC
X SYSTEM taken from the internal library. A system containing both library and
SPLIT characterized/user defined components should contain both LNAMES and CNAMES
GROUP with the default specification, 1*, where appropriate.
COMB
SIMULATE The component names are translated into upper case on input. The length may be up
REGRESS
BLACKOIL
to 72 characters, but a limit of four is suggested to fit into the program output formats.
PSEUCOMP Each record is terminated with a slash (/).
OUTECL3
VFP
Example
For a nine-component condensate system, with the first five from library and the last
four user-defined:
CNAMES
5* C4-6 C7+1 C7+2 C7+3 /
LNAMES
CO2 N2 C1 C2 C3 4* /
Keywords 227
LNAMES
MAXIT Max. number of regression iterations
Specifies maximum number of function evaluations in the non-linear regression step.
RUNSPEC
SYSTEM This does not include the initial stage required in setting up the numerical Jacobian.
SPLIT
GROUP
The default is 50.
COMB The record is terminated with a slash (/).
SIMULATE
X REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
MAXIT
40 /
228 Keywords
MAXIT
MAXSTEP Maximum step size allowed in regression
At each step of the regression, a maximum step size, or trust region radius is
RUNSPEC
SYSTEM maintained. This is allowed to vary during the regression, depending upon how
SPLIT successfully the behavior of the regression function is predicted. However, a maximum
GROUP trust region radius is also imposed. MAXSTEP allows this to be reset.
COMB
SIMULATE The variables upon which this limit acts are scaled to unity in the case of
X REGRESS
BLACKOIL p c , T c , V c , Z c , Ω a , Ω b , and are the actual values for binary interaction coefficients and
PSEUCOMP acentric factors. Thus a maximum step of 0.1 corresponds to a change of 10% in a
OUTECL3
VFP critical temperature.
The default is 0.1.
The record is terminated with a slash (/).
Example
MAXSTEP
0.2 /
Keywords 229
MAXSTEP
MDP Data for Whitson splitting
Specifies the α (shape factor) and η (lowest mole weight) parameters used in the
RUNSPEC
SYSTEM Whitson probability density function used to characterize plus fractions.
X SPLIT
GROUP Default values are 1.0 and 97.0 respectively.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
Change η to 92.0, leave α unchanged:
MDP
1* 92.0 /
230 Keywords
MDP
MESSAGE Echo message to file and screen
Echoes a single line to the screen and print file when a PVI input file is read. The
X RUNSPEC
X SYSTEM MESSAGE keyword can occur anywhere after the first keyword, RUNSPEC, and takes
X SPLIT the form:
X GROUP
X COMB
X SIMULATE
X
X
REGRESS
BLACKOIL Example
X PSEUCOMP
X OUTECL3
X VFP MESSAGE
Message to echo goes here
Keywords 231
MESSAGE
MINDELP Minimum pressure difference
Specifies the minimum pressure step to allow during the automatic insertion of
RUNSPEC
SYSTEM pressure nodes into pressure depletion experiments following blackoil table total
SPLIT compressibility tests.
GROUP
COMB
SIMULATE Note This keyword must be placed after the keywords defining the depletion and
REGRESS separator experiments.
X BLACKOIL
X PSEUCOMP
OUTECL3
VFP
The default is one atmosphere.
Example
MINDELP
14.7 /
232 Keywords
MINDELP
MINSTEP Minimum step limit allowed in regression
At each step of the regression, a maximum step size, or trust region radius is
RUNSPEC
SYSTEM maintained. This is allowed to vary during the regression, and may be reduced
SPLIT
GROUP
• if the function is poorly predicted
COMB • if error conditions occur in the regression functional
SIMULATE
X REGRESS • if the solution attempts to cross a lower or upper limit.
BLACKOIL
PSEUCOMP If the step size limit falls below a minimum value, for any reason, the regression
OUTECL3
VFP
terminates, as further progress is unlikely to be made. This minimum value may be
reset with the MINSTEP keyword.
Note This is not the minimum step that may be taken; it is the minimum upper limit
imposed on a step.
The default is 0.00001.

Example
MINSTEP
0.000001 /
Keywords 233
MINSTEP
MIX Mix samples
This keyword allows two samples to be mixed to form a new one. The inputs are the
RUNSPEC
X SYSTEM two sample names, the new sample name, the type of mixing and the amount of
SPLIT mixing. The types of mixes are as follows:
GROUP
COMB • MOL specifies that the mixing amount given is the mole fraction of sample two in
SIMULATE the mixture with sample 1.
REGRESS
BLACKOIL • If the mixing method is GOR, then the amount is given as the gas volume at
PSEUCOMP standard conditions of sample two as a ratio of the volume of the initial sample.
OUTECL3
VFP The latter is useful for consideration of mixing of lean gases in oil samples.
Note Mixing is only performed if:

- the two samples are different
- the amount to mix is greater than zero
- the name of the new sample is different from any other sample
- the mixing does not take the number of samples over the program limit.
• For the GOR option, the volume of the initial sample is usually the volume at its
P sat at the specified temperature. However, an alternative pressure can be
specified; this is shown in Example 2, below.
Each record must be terminated with a slash (/).
The set of records must end with a blank record, containing only a slash (/).
Examples
Example 1
Mix sample Z1 with Z2 to form Z3. The amount to mix is such that sample Z2 is 0.45 of
the mole fraction in the new sample Z3.
MIX
Z1 Z2 Z3 80.000 MOL 0.45 /
/
Example 2
Mix sample T1 with T2 to form T3. The amount to mix is represented as a GOR of 5.0
MSCF/stb, and the volume of the first sample T1 is measured at a pressure of 5000 PSI
at the mix temperature of 670 °F . If the saturated volume of T1 was required then the
PRES could be set as 0.0, or those two items removed from the keyword.
MIX
T1 T2 T3 670.0 PRES 5000.0 GOR 5.0 /
/
234 Keywords
MIX
MODSPEC Denotes start of the run specification section
Any split, group, regress or user change of the fluid system causes a new
x RUNSPEC
SYSTEM RUNSPEC/SYSTEM section to be appended to the audit trail, which can be saved on exit
SPLIT from PVTi to a PVI file. These new sections are labelled with the keywords MODSPEC
GROUP and MODSYS to indicate they are modified definitions. They are interchangeable with
COMB
SIMULATE RUNSPEC and SYSTEM.
REGRESS
BLACKOIL
This section defines the number of components, equation of state and viscosity option,
PSEUCOMP run title and name of any new output file.
OUTECL3
VFP The keyword MODSPEC has no arguments.
Keywords 235
MODSPEC
MODSYS Start of the MODSYS section
MODSYS is a delimiter keyword, specifying the start of the MODSYS section, which
RUNSPEC
X SYSTEM defines the component properties and the fluid sample compositions.
SPLIT
GROUP
As for MODSPEC, the MODSYS keyword denotes a modified fluid system that was saved
COMB from a previous PVTi session following on from a split, group, regress or manual
SIMULATE change of the fluid system originally defined with RUNSPEC and SYSTEM. The MODSYS
REGRESS
BLACKOIL
section generally follows the MODSPEC section.
PSEUCOMP The keyword is interchangeable with SYSTEM.
OUTECL3
VFP
236 Keywords
MODSYS
MOSES Blackoil tables
Specifies that the Moses method for the generation of composite from differential
RUNSPEC
SYSTEM blackoil tables is used. By default, the program uses the Whitson and Torp method of
SPLIT composite black oil table generation.
GROUP
COMB The keyword MOSES has no arguments.
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 237
MOSES
MW Specify molecular weights
Specifies the molecular weights for the components in the system. The keyword is
RUNSPEC
SPLIT RUNSPEC section.
GROUP
COMB The set of records must be terminated with a forward slash (/) character.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
Molecular weights for 15-component system:
MW
44.010 28.013 16.043 30.070 44.097 58.124
58.124 72.151 72.151 86.178 106.09 152.68
209.26 281.29 462.30 /
238 Keywords
MW
MWS Define plus fraction mole weight for CMF
splitting
This keyword specifies the mole weight of the plus fraction in the Constant Mole
RUNSPEC
SYSTEM
Fraction splitting algorithm. The keyword is followed by a single real number for the
X SPLIT mole weight and a forward slash (/) character.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL Example
PSEUCOMP
OUTECL3 Plus fraction mole weight of 140.0:
VFP
MWS
140.0 /
Keywords 239
MWS
NCOMPS Specify number of components
Specifies the number of components in the sample. This keyword must be entered in
X RUNSPEC
SYSTEM the RUNSPEC section. The maximum value available for NCOMPS is usually 100. This is
SPLIT set by a parameter in the source code, and can be increased at specific request.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL Example
PSEUCOMP
OUTECL3
VFP NCOMPS
20 /
240 Keywords
NCOMPS
NEWPVI Request new output PVI file
Allows you to output the current system description at the end of a batch run if it is
X RUNSPEC
SYSTEM thought that the system changes during the run (because of splitting/grouping and/or
SPLIT regression). The keyword is followed by the required filename root (that is the file
GROUP name less any prefix or suffix) and is terminated with a forward slash (/) character.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP Example
OUTECL3
VFP
Write out new PVI file with name SAVE:
NEWPVI
SAVE /
Keywords 241
NEWPVI
NEWPVO Request new output PVO file
Allows you to name the export file produced in batch mode when there is an OUTECL3
RUNSPEC
SYSTEM section present. The keyword is followed by the required filename root (that is the file
SPLIT name less any prefix or suffix) and is terminated with a forward slash (/) character.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL Example
PSEUCOMP
X OUTECL3 Write a PVO file with name SAVE:
VFP
NEWPVO
SAVE /
242 Keywords
NEWPVI
NOECHO No insertion of PVI file into PVP file
This keyword requests that the PVI input file is not to be inserted into the PVP output
X RUNSPEC
SYSTEM file. This is the default condition.
SPLIT
GROUP
The keyword NOECHO has no arguments.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 243
NOECHO
OBS Specify observations
Specifies the observations associated with experiments performed individually or as
RUNSPEC
SYSTEM part of a regression analysis.
SPLIT
GROUP
X COMB Note This keyword can only be specified if the EXP keyword and associated
X SIMULATE experiments have already been defined.
X REGRESS
BLACKOIL
PSEUCOMP • In the SIMULATE section, this keyword provides the means of supplying
OUTECL3
VFP
experimentally determined data for comparison purposes in output.
• In the REGRESS section, this keyword provides the experimental data to be
regressed against.
The form of the keyword and its associated data is:
OBS
no mnem weight data /
mnem weight data /
:
/
no mnem weight data /
mnem weight data /
:
/
/
where:
• no specifies the experiment number given in the EXP keyword

• mnem specifies the observation mnemonic (see below)
• weight global weight to be given to subsequent data.
Several observations can be associated with a given experiment, each of which must be
terminated with a forward slash (/) character. The entry of observations for a particular
experiment is then terminated with a forward slash (/) character. A forward slash
character (/) also terminates the entry of all observations.
244 Keywords
OBS
OBSIND Specify observation weights
Specifies the individual observation weights associated with observations defined
RUNSPEC
SYSTEM with an OBS keyword.
SPLIT
GROUP
X COMB Note Note that the OBS keyword must exist in the section of the PVI file before the
X SIMULATE OBSIND keyword is inserted.
X REGRESS
BLACKOIL
PSEUCOMP If no OBSIND keyword is used the individual observation weights take the value of the
OUTECL3
VFP
global observation weight defined for each mnemonic in OBS.
The form of the keyword and its associated data is:
OBSIND
no mnem data /
mnem data /
:
/
no mnem data /
mnem data /
:
/
/
where:
• no specifies the experiment number given in the EXP keyword (page

205);
• mnem specifies the observation mnemonic (see OBS keyword, page 244);
• data individual weights to be given to each point described in the previous
OBS keyword.
Example
Observation weights for the experiments defined for the EXP keyword
(compare with example in OBS keyword description):
1 Saturation pressure and vapor Z -factor of dew point
2 Saturation pressure and swelling factor (relative volume) at each stage of swelling
test
3 Relative volumes at each stage of CCE, vapor Z -factors at pressures above
saturation pressure only
4 Liquid saturation, vapor Z -factor and gas mole fractions at each stage of CVD
(nine component system)
5 GOR at each stage and stock tank GOR (in Mscf/stb) from two stage separators.
Keywords 245
OBSIND
OBSIND
1 PSAT 50.0 /
ZV 1.0 /
/
2 PSAT 10.0 10.0 10.0 10.0 /
RELV 1.0 1.0 1.0 1.0 /
/
3 RELV 1.0 1.0 1.0 1.0 1.0 1.0
1.0 1.0 1.0 1.0 1.0 1.0
2.0 2.0 3.0 4.0 /
ZV 1.0 1.0 1.0 1.0 1.0 1.0 10* /
/
4 SL 1.0 1.0 1.0 1.0 /
ZV 1.0 1.0 1.0 1.0 /
YMF 1.0 1.0 1.0 1.0
1.0 1.0 1.0 1.0 1.0
1.0 1.0 1.0 1.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
1.0 1.0 1.0 1.0 1.0 /
/
5 GOR 100.0 10.0 /
TGOR 10.0 /
/
/
246 Keywords
OBSIND
OMEGAA,OMEGAB
Specify EoS omega values
These keywords specify non-default Ω a and Ω b values for the components in the
RUNSPEC
X SYSTEM system. Each keyword is followed by N c values, where N c is the number of
SPLIT
GROUP components in the system, and is terminated with a forward slash (/) character. You
COMB need only specify these keywords if you wish to override the default equation of state
SIMULATE
REGRESS
values, which are as follows:
BLACKOIL
PSEUCOMP
OUTECL3 Table 7.7 Equation of State omega values
VFP
Equation OMEGAA OMEGAB
RK 0.4274802 0.08664035
SRK 0.4274802 0.08664035
ZJ 0.4274802 0.08664035
PR 0.457235529 0.07776074
Example
Ω a and Ω b values for two component system:
OMEGAA
0.457 0.456 /
OMEGAB
0.0780 0.0781 /
Keywords 247
OMEGAA, OMEGAB
OPTIONS Set various program options
This keyword is used to set the various program options. Each option is described in
X RUNSPEC
SYSTEM "Program options" on page 156. As from the release of PVTi 2004A there are 21 options
SPLIT with each option specified by an integer (usually 0 or 1).
GROUP
COMB
SIMULATE Note Upon program start-up, most options are OFF, specified by an option value 0
REGRESS (zero). Options 2 and 7 are the only ones that are on (=1) by default.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
To subsequently turn any option ON, specify an the integer 1 (but be careful of the
special options below). The vector of integers is terminated with a / character.
Special option settings

Option 4
Temperature dependence of volume shifts (when using a three-parameter EoS) takes
the following arguments:
• 0 - No temperature dependence;
• 1 - Apply linear thermal expansion to all components;
• 2 - Use polynomial correlations and thermal expansion.
Option 11
Optional printing of results to the PVP file also takes specific arguments:
• 0 - ALways print;
• 1 - OPTionally print (that is user prompted);
• 2 - NEVer print.
Option 14
Alternative definitions of GOR in differential liberation experiments also takes specific
arguments:
• 0 - Default definition;
• 1 - Last stage removed and oil volume at its bubble point pressure;
• 2 - Incremental GOR;
• 3 - As default but volume of oil at its bubble point pressure rather than stock tank
conditions.
248 Keywords
OPTIONS
Example
Separator liquid volumes to be output at stock tank conditions. Cheuh-Prausnitz BICs
required. Optional printing of results to PVP file:
OPTIONS
0 0 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 /
Keywords 249
OPTIONS
OUTECL3 Start of the OUTECL3 section
This is a delimiter keyword, specifying the start of the OUTECL3 section.
RUNSPEC
SYSTEM
GROUP
COMB
SIMULATE
REGRESS
The section is used in generating output data for inclusion into the ECLIPSE
BLACKOIL compositional model. Either equation of state or pressure-dependent tabular data can
PSEUCOMP be generated.
X OUTECL3
VFP
250 Keywords
OUTECL3
Keywords P- S
This section contains the P-S keywords.
• "PARACHOR Define parachors" on page 252
• "PCRIT Critical pressures" on page 253
• "PEARCE Blackoil tables" on page 254
• "PEDERSEN Specify Pedersen viscosities" on page 255
• "PRCORR Peng-Robinson correction" on page 256
• "PSEUCOMP Start of the PSEUCOMP section" on page 257
• "RECOVERY Liquid production for recovery estimates" on page 258
• "REGRESS Start of the REGRESS section" on page 259
• "REGTARG Regression target" on page 260
• "RTEMP Reservoir temperature for ECLIPSE Compositional" on page 261
• "RUNSPEC Denotes start of the run specification" on page 262
• "SALINITY Specify sample salinity" on page 263
• "SAMPLE Specify fluid sample" on page 264
• "SAMPLES Specify fluid samples" on page 265
• "SAMPLES Specify fluid samples" on page 266
• "SAMTITLE Specify titles of fluid samples" on page 267
• "SAVCOMP Save compositions" on page 268
• "SCT Defines Semi-Continuous Thermodynamics split" on page 269
• "SG Specify specific gravity" on page 270
• "SIMULATE Start of the SIMULATE section" on page 271
• "SPECHA,B,C,D Specify specific heat capacity coefficients" on page 272
• "SPLIT Start of the SPLIT section" on page 273
• "SSHIFT Dimensionless volume shifts for PR3" on page 274
• "STCOND Standard conditions" on page 275
• "SYSTEM Start of the SYSTEM section" on page 276

Keywords 251
Keywords P- T
PARACHOR Define parachors
Specifies the parachors for the components in the fluid system. The keyword is
RUNSPEC
X SYSTEM followed by up to N c values (the number of components specified in the RUNSPEC
SPLIT section) and terminated by a forward slash (/) character.
GROUP
COMB
SIMULATE
REGRESS Units
BLACKOIL
PSEUCOMP Should be consistent to obtain surface tensions in dyne/cm.
OUTECL3
VFP
Example
Parachors from [Ref. 18] for N2, C1, C2 and C3:
PARACHOR
41.0 77.0 108.0 150.3 /
252 Keywords
PARACHOR
PCRIT Critical pressures
Specifies the critical pressures for the components in the system. The keyword is
RUNSPEC
X SYSTEM followed by N c (the number of components in the system) values, and terminated with
SPLIT a forward slash (/) character.
GROUP
COMB • UNITS: bars (METRIC), psi (FIELD), atm (LAB), atm (PVT-M).
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
For a two-component system, using atm:
PCRIT
72.90 36.00 /
Keywords 253
PCRIT
PEARCE Blackoil tables
Specifies that the Pearce method for generating blackoil tables is used in preference to
RUNSPEC
SYSTEM the Whitson and Torp or Coats method. By default, the program uses the Whitson and
SPLIT Torp method.
GROUP
X COMB The keyword PEARCE has no arguments.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
254 Keywords
PEARCE
PEDERSEN Specify Pedersen viscosities
Specifies that the Pedersen corresponding state viscosity correlation is to be used,
X RUNSPEC
SYSTEM rather than the default Lohrenz-Bray-Clark method.
SPLIT
GROUP
The keyword PEDERSEN has no arguments.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Keywords 255
PEDERSEN
PRCORR Peng-Robinson correction
This keyword requests that a modified form of Peng-Robinson equation of state is to
X RUNSPEC
SYSTEM be used. This changes the usual form of the Ω a value as a function of the component
SPLIT acentric factor.
GROUP
COMB The keyword PRCORR takes no arguments, and has no effect on equations of state other
SIMULATE
REGRESS than the Peng-Robinson.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
256 Keywords
PRCORR
PSEUCOMP Start of the PSEUCOMP section
This is the delimiter keyword specifying the start of the PSEUCOMP section used for the
RUNSPEC
SYSTEM generation of ECLIPSE pseudo-compositional tables.
SPLIT
GROUP
This section performs a depletion experiment to define fluid properties in the reservoir
COMB and then passes the liquid and vapor through a separator network to surface
SIMULATE conditions to define standard blackoil tables. Then a series of gas injections are
REGRESS
BLACKOIL
performed and modified blackoil tables generated with the various mixtures.
X PSEUCOMP
OUTECL3
VFP
Keywords 257
PSEUCOMP
RECOVERY Liquid production for recovery estimates
The presence of the RECOVERY keyword turns on the ability to calculate liquid
RUNSPEC
SYSTEM production by the method of Reudelhuber and Hinds ([Ref. 37]) in the COMB section
SPLIT recovery calculations. It should be followed by the coordinates of two points on the
GROUP curve of relative permeability versus total liquid saturation. The order of entry is the
X COMB
SIMULATE minimum and maximum relative permeabilities, and then the total liquid saturations
REGRESS for the minimum and maximum points.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
RECOVERY
<RPMIN> <RPMAX> <SRPMIN> <SRPMAX> /
258 Keywords
RECOVERY
REGRESS Start of the REGRESS section
This is a delimiter keyword, specifying the start of the REGRESS section.
RUNSPEC
SYSTEM
GROUP
COMB
SIMULATE
X REGRESS
This section is used to perform a regression of the equation of state parameters such as
BLACKOIL to minimize the difference between observed and calculated results of the following
PSEUCOMP experiments:
OUTECL3
VFP • equilibrium flash
• bubble and dew point evaluation
• saturated pressure
• constant composition expansion
• constant volume depletion
• differential liberation
• swelling test
• multi-stage separator networks
• the multi-contact test.
Keywords 259
REGRESS
REGTARG Regression target
This quantity allows the default regression target to be changed. The regression target
RUNSPEC
SYSTEM is applied to the sum of squares of the differences between the calculated and observed
SPLIT values. If a solution is found for which the residual approaches zero, the regression
GROUP returns successfully when the sum of squares falls below this value. If there is a non-
COMB
SIMULATE zero residual minimum this limit is applied to the derivatives of the residual at the
X REGRESS minimum.
BLACKOIL
PSEUCOMP • DEFAULT: 0.000001.
OUTECL3
VFP The record is terminated by a forward slash (/) character.
Example
RETARG
0.1D-8 /
260 Keywords
REGTARG
RTEMP Reservoir temperature for ECLIPSE
Compositional
This keyword defines the reservoir temperature to be used in an ECLIPSE
RUNSPEC
SYSTEM
Compositional simulation by including it along with the EoS data, which is output on
SPLIT the PVO file. It is followed by a single real number and terminated with a forward slash
GROUP (/) character. It should be in the current PVTi unit set: it may be converted on output
COMB
SIMULATE
into an appropriate ECLIPSE Compositional unit set.
REGRESS
BLACKOIL
PSEUCOMP
X OUTECL3
VFP
Example
Specify a reservoir temperature of 200 °F :
RTEMP
200.0 /
Keywords 261
RTEMP
RUNSPEC Denotes start of the run specification
This keyword opens the run specification section, and is normally the first keyword in
X RUNSPEC
SYSTEM the input data file. This section defines the number of components, equation of state
SPLIT and viscosity option, run title and name of any new output file.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
262 Keywords
RUNSPEC
SALINITY Specify sample salinity
For systems containing water, this specifies the salinity of any sample. The data are
RUNSPEC
X SYSTEM only used in the multiphase flash experiment.
SPLIT
GROUP
The record is terminated by a forward slash (/) character.
COMB • UNITS: ppm (parts per million).
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
RTEMP
200.0 /
Keywords 263
SALINITY
SAMPLE Specify fluid sample
This keyword is specifies the fluid sample to use in the exportation of a fluid model for
RUNSPEC
SYSTEM use in an ECLIPSE Compositional simulation.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
X OUTECL3
VFP
APITRACK
264 Keywords
SAMPLES
SAMPLES Specify fluid samples
This keyword defines different composition fluids. Up to 50 different samples can be
RUNSPEC
X SYSTEM defined on all platform types. This maximum number can be changed on request,
SPLIT please contact us for advice if this situation arises. Each composition must be given a
GROUP unique mnemonic to distinguish it from all others.
COMB
SIMULATE
REGRESS Note Note that the default sample composition is given by the ZI keyword and is
BLACKOIL
PSEUCOMP given the mnemonic of ZI.
OUTECL3
VFP
Thus, each line consists of the composition mnemonic followed by up to N c mole
fractions, which should sum to unity, followed by a forward slash (/) character. The
last composition should be followed by another forward slash (/) character.
The composition of the lean gas injection fluid used in a swelling test should be defined
under this keyword.
Example
Lean gas composition for swelling test:
SAMPLES
ZINJ 2* 0.9468 0.0527 0.0005 4* /
/
Keywords 265
SAMPLES
SAMPLES Specify fluid samples
If this keyword is in the APITRACK section and not the SYSTEM section then it specifies
RUNSPEC
SYSTEM the fluid samples involved when exporting a series of black oil tables to be used with
SPLIT the API Tracking option in ECLIPSE BlackOil.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
266 Keywords
SAMPLES
SAMTITLE Specify titles of fluid samples
This keyword defines a long description to be associated with each alternative sample,
RUNSPEC
X SYSTEM so as to give more information about the origin of the samples, etc. The description
SPLIT must be enclosed in quotes.
GROUP
COMB Each record is terminated by a forward slash (/) character. The set of records must be
SIMULATE ended by a blank record, containing only a slash.
REGRESS
BLACKOIL
PSEUCOMP Note This keyword should not precede the SAMPLES keyword.
OUTECL3
VFP
Example
Second sample from deeper in the column. Lean gas composition for swelling test:
SAMTITLE
Z2 ‘Second sample from deeper in the column’ /
ZINJ ‘Lean gas composition for swelling test’ /
/
Keywords 267
SAMTITLE
SAVCOMP Save compositions
If you have turned on the OPTIONS switch for saving compositions calculated during
RUNSPEC
SYSTEM experiments, this keyword may be used in the SIMULATE section to save compositions
SPLIT to named samples. Each line in this section contains the experiment number, the fluid
GROUP type required (from XMF, YMF, ZMF, for liquid, vapor and total compositions), the stage
COMB
X SIMULATE of the experiment if it is a multi-stage experiment, and the sample name required. The
REGRESS stage index should be omitted if the experiment is not a multi-stage experiment. In the
BLACKOIL particular case of the stock tank stage of a SEPS experiment, you should enter ST for
PSEUCOMP
OUTECL3 the stage indicator.
VFP
Each record is terminated by a forward slash (/) character. The set of records must be
ended by a blank record, containing only a slash.
Example
Save the total composition for experiment 4, a COMPG experiment, at stage (here depth)
4, and label it Z80. Save the liquid output from stage 3 of experiment 2, a SEPS
experiment, and label it S23, and the stock tank vapor output of the same experiment,
labelling it ST2:
SAVCOMP
4 ZMF 4 Z80 /
2 XMF 3 S23 /
2 YMF ST ST2 /
/
268 Keywords
SAVCOMP
SCT Defines Semi-Continuous Thermodynamics
split
This keyword requests the modified Whitson method, here denoted the Semi-
RUNSPEC
X SYSTEM
Continuous-Thermodynamics (SCT) split. This method has the advantage that it can
X SPLIT take multiple plus fraction definitions from multiple samples and split them into a
GROUP consistent set of pseudo-components with mole fractions adjusted between the
COMB
SIMULATE
different samples to honour moles and mole weights.
REGRESS
Each record is terminated by a forward slash (/) character. The set of records must be
BLACKOIL
PSEUCOMP ended by a blank record, containing only a slash.
OUTECL3
VFP The keyword is followed by N samp + 1 lines of information, where N samp is the
number of already defined samples, including the default sample ZI.
• The first line of data specifies the mnemonic of the heavy end to split, the number
of pseudo-components required (which must be between two and five), the
minimum mole weight in the plus fraction (the Whitson η -parameter) and the
mole weight of the heaviest pseudo-component required. As a default, the heaviest
pseudo-component weight could be set to 1.5 and 2.0 times the heaviest plus
fraction mole weight.
• Next follows N samp lines of data, one for each sample currently defined. On each
line should be the sample mnemonic, the skewness parameter (Whitson α -
parameter), the sample plus fraction mole weight and specific gravity.
Example
The default and one alternative sample from different depths in the hydrocarbon
column, and therefore different mole weights and specific gravities for the C7+ plus
fraction. Different compositions are on the ZI and SAMPLES keyword, sample ZALT,
which preceded this keyword and data. Three pseudo-component split requested with
minimum and maximum mole weights of 90.0 and 300.0. Different skewness
parameters reflects compositional grading in fluid from top (ZI - low Alpha) to bottom
(ZALT - high Alpha).
SCT
C7+ 3 90.0 300.0 /
ZI 1.10 130.0 0.75 /
ZALT 2.30 150.0 0.80 /
/
Keywords 269
SCT
SG Specify specific gravity
Specifies the specific gravity (with respect to water at standard conditions) for the plus
RUNSPEC
SYSTEM fraction to be split. The record is terminated by a forward slash (/) character.
Z SPLIT
GROUP
COMB
SIMULATE
REGRESS
Example
BLACKOIL
PSEUCOMP SG
OUTECL3
VFP 0.86 /
270 Keywords
SG
SIMULATE Start of the SIMULATE section
This is a delimiter keyword, specifying the start of the SIMULATE section.
RUNSPEC
SYSTEM
GROUP
COMB
X SIMULATE
REGRESS
This section is used to define any of the following experiments:
BLACKOIL • equilibrium flash
PSEUCOMP
OUTECL3 • bubble and dew point evaluation
VFP
• constant composition expansion
• constant volume depletion
• differential liberation
• swelling test
• multi-stage separator networks
• multiphase flashes
• ternary diagrams
• saturated pressure and temperature
• critical point
• compositional gradient
• minimum miscibility pressures
all using the equation of state model.
Keywords 271
SIMULATE
SPECHA,B,C,D
Specify specific heat capacity coefficients
Specifies the specific heat capacity coefficients for the components in the system. The
RUNSPEC
X SYSTEM keywords are followed by N c values, and terminated with a / character, where N c is
SPLIT the number of components specified in the RUNSPEC section.
GROUP
COMB
SIMULATE
REGRESS Note The values are output only in EoS data for ECLIPSE Compositional input files
BLACKOIL and only if OPTION number 7 is turned on.
PSEUCOMP
OUTECL3
VFP
Example
For a two component system, in PVT-Metric units:
-- Specific Heat Coefficients A kJ/kgm/K

SPECHA
5.409000000E+00 -4.224000000E+00 /
-- Specific Heat Coefficients B kJ/kgm/K
SPECHB
1.781000000E-01 3.063000000E-01 /
-- Specific Heat Coefficients C kJ/kgm/K
SPECHC
-6.938000000E-05 -1.586000000E-04 /
-- Specific Heat Coefficients D kJ/kgm/K
SPECHD
8.713000000E-09 3.215000000E-08 /
272 Keywords
SPECHA,B,C,D
SPLIT Start of the SPLIT section
This is a delimiter keyword, specifying the start of the SPLIT section.
RUNSPEC
SYSTEM
X SPLIT Note If present, this section must not appear before the SYSTEM section.
GROUP
COMB
SIMULATE
REGRESS
This section is used to define data necessary for the splitting of the plus fraction,
BLACKOIL assumed to be the last component.
PSEUCOMP
OUTECL3
VFP
Keywords 273
SPLIT
SSHIFT Dimensionless volume shifts for PR3
This keyword defines the dimensionless volume shift parameters used in the three-
RUNSPEC
X SYSTEM parameter Peng-Robinson equation of state. The keyword is followed by N c real
SPLIT
GROUP
numbers and terminated with a / character, where N c is the number of components.
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP Example
OUTECL3
VFP
SSHIFT
0.15 0.05 -0.05 -0.03 -0.01 0.01 0.05 0.12 /
274 Keywords
SSHIFT
STCOND Standard conditions
Specifies the current standard conditions. This is followed by the standard temperature
RUNSPEC
X SYSTEM and pressure values. The record is terminated by a forward slash (/) character.
SPLIT
GROUP
COMB
SIMULATE
REGRESS
Example
BLACKOIL
PSEUCOMP Standard conditions in degrees Fahrenheit and psia:
OUTECL3
VFP
STCOND
60.0 14.7 /
Keywords 275
STCOND
SYSTEM Start of the SYSTEM section
This is a delimiter keyword, specifying the start of the SYSTEM section. This section
RUNSPEC
X SYSTEM generally follows the RUNSPEC section.
SPLIT
GROUP
This section is used to define the component properties and the fluid sample
COMB compositions.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
276 Keywords
SYSTEM
Keywords T - Z
This section contains the T-Z keywords.
• "TBOIL Specify boiling points" on page 278
• "TCRIT Specify critical temperatures" on page 279
• "THERMX Thermal expansion coefficient for volume shifts" on page 280
• "TITLE Specify run title" on page 281
• "TLOW Define lowest temperature for VFP tables" on page 282
• "TREF Specify reference temperatures" on page 283
• "UNITS Specify unit conventions" on page 284
• "VAR Specify regression variables" on page 285
• "VCRIT Specify volumes" on page 289
• "VCRITVIS Specify volumes for LBC viscosity calculations" on page 290
• "VFP Start of the VFP section" on page 292
• "WAT100 Output water properties" on page 293
• "WATVFP Output water properties" on page 296
• "WETGAS Wet gas tables" on page 297
• "WHIT Defines Whitson splitting" on page 298
• "WHITSON Blackoil tables" on page 299
• "XMFVP/YMFVP XMFVP and YMFVP ECLIPSE tables" on page 300
• "ZCRIT Specify critical Z-factors" on page 301
• "ZCRITVIS Specify critical Z-factors for LBC calculations" on page 302
• "ZI Specify sample composition" on page 303
• "ZMFVD Composition versus depth table" on page 304

"Keywords P- S" on page 251.
Keywords 277
Keywords U- Z
TBOIL Specify boiling points
Specifies the boiling point temperatures for the components in the system. The
RUNSPEC
X SYSTEM keyword is followed by N c values, where N c is the number of components specified in
SPLIT the RUNSPEC section, and terminated with a / character. The current temperature
GROUP
COMB convention is used. The default is degrees K , and for alternatives a DEGREES keyword
SIMULATE should have been previously read.
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3 Note Boiling points are only required if the Zudkevitch-Joffe equation has been
VFP selected.
Example
For a two component CO2/I-C4 system, using degrees K :
TBOIL
194.70 261.30 /
278 Keywords
TBOIL
TCRIT Specify critical temperatures
Specifies the critical temperatures for the components in the system. The keyword is
RUNSPEC
SPLIT RUNSPEC section, and terminated with a forward slash (/) character. The current
GROUP
COMB temperature convention is used.
SIMULATE
REGRESS • DEFAULT: degrees K . For alternatives a DEGREES keyword should have been
BLACKOIL previously read.
PSEUCOMP
OUTECL3
VFP
Example
TCRIT
305.6 408.1 /
Keywords 279
TCRIT
THERMX Thermal expansion coefficient for volume
shifts
Specifies the value of the volume expansion coefficient which may be applied to the
RUNSPEC
X SYSTEM
three-parameter equations of state (PR3, SRK3) volume shifts.
SPLIT
It is terminated with a forward slash (/) character.
GROUP
COMB
SIMULATE
REGRESS Note The specified value is only used if the appropriate OPTIONS switch is in force.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
–1
Default value with temperature in degrees Centigrade, that is THERMX in °C :
THERMX
0.0005 /
280 Keywords
THERMX
TITLE Specify run title
This keyword introduces the title of the run, which appears on the header of each
X RUNSPEC
SYSTEM section of the PVP output file. The line following the title keyword is taken as the title
SPLIT to be used.
GROUP
COMB
SIMULATE Note The syntax is slightly different from that of other keywords, in that no
REGRESS terminating forward slash (/) character is required.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
Example
TITLE
18 component condensate test
Keywords 281
TITLE
TLOW Define lowest temperature for VFP tables
PVT properties for VFPi must be defined at least for two temperatures. The highest
RUNSPEC
SYSTEM temperature, assumed to be reservoir temperature, is set on the constant composition
SPLIT expansion (CCE) experiment definition under the EXP keyword. The lowest
GROUP temperature in the production string is defined using this keyword.
COMB
SIMULATE It is terminated with a forward slash (/) character.
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
X VFP Example
Production string runs through deep sea-water at 40 °F
TLOW
40.0 /
282 Keywords
TLOW
TREF Specify reference temperatures
Specifies the reference temperatures for the components in the system. The keyword is
RUNSPEC
SPLIT RUNSPEC section, and terminated with a forward slash (/) character. The current
GROUP
COMB temperature convention is used.
SIMULATE
REGRESS • DEFAULT: degrees K . For alternatives a DEGREES keyword should have been
BLACKOIL previously read.
PSEUCOMP
OUTECL3
VFP
Note These reference temperatures are only required if the Zudkevitch-Joffe
equation has been selected.
Example
TLOW
40.0 /
Keywords 283
TREF
UNITS Specify unit conventions
This keyword specifies which unit convention is required, and sets flags to specify
RUNSPEC
X SYSTEM whether absolute or gauge pressures, and mole fractions or mole percentages, are to be
SPLIT used.
GROUP
COMB The UNITS keyword is followed by a character string identifying the units convention,
SIMULATE which may be one of the following four:
REGRESS
X BLACKOIL
PSEUCOMP
X OUTECL3
VFP
X APITRACK
METRIC: Metric unit system

FIELD: Field units
LAB: Laboratory units
PVT: PVT-metric units (metric units with pressures in atm).
Only the first character is significant, and case is not significant.
• DEFAULT: PVT-metric.
The pressure switch is one of the two following strings:
ABSOL: Absolute pressure
GAUGE: Gauge pressure.
• DEFAULT: ABSOLute pressures.
The mole composition switch is one of the following strings:
FRAC: Compositions as fractions of unity
PERC: Compositions as percentages of a hundred.
• DEFAULT: FRACtions.
Example
Set to field units:
UNITS
F /
284 Keywords
UNITS
VAR Specify regression variables
Specify variables for regression. The keyword is followed by a number of lines of data,
RUNSPEC
SYSTEM each specifying variables to be added to the regression set. The data is terminated by a
SPLIT blank record (a single forward slash (/) character).
GROUP
COMB The keyword thus takes the form:
SIMULATE
X REGRESS
BLACKOIL VAR
PSEUCOMP S1 I1 J1 K1 L1 IND1LL1 UL1/
OUTECL3 S2 I2 J2 K2 L2 IND2LL2 UL2 /
VFP S3 I3 J3 K3 L3 IND3LL3 UL3 /
: : : : : : : :
/
• The strings S1,S2... identify variables for regression, the indices I1, J1 (and K1, L1
for binaries) identify component ranges, lower and upper.
Possible strings identifying variables (that is, S1, S2 etc.) are:
• TCRIT
Critical temperatures
• PCRIT
Critical pressures
• VCRIT
Critical volumes in LBC viscosity correlation
• ZCRIT
Critical Z -factors in LBC viscosity correlation
• ACF
Acentric factors
• OMEGAA
Ω a values
• OMEGAB
Ω b values
• SSHIFT
Volume shift parameter
• BIC
Binary interaction coefficients
In all cases except BIC, the lower and upper component indices should be
specified. In the case of BIC, lower and upper index ranges should be given, that
is, in four values. For example BIC 1 3 7 8 contains the BICS for (1, 7), (1, 8), (2, 7),
(2, 8), (3, 7) and (3, 8).
• The indices IND1, IND2, etc., specify the variable index as applied to the particular
variable.
Keywords 285
VAR
“Special” regression variables are available and are:
• SCTALF
SCT Splitting parameter
• SCTMW
SCT plus fraction weight
• SCTKW
SCT Watson K factor
• THERMX
Thermal expansion coefficient
• CHEUHA
Cheuh-Prausnitz BIC coefficient
• CHARMW
Molecular weights of characterized components
• CHARMF
Mole Fractions of characterized and/or user components for any sample
• MIXING
The mixing factor between two samples when using the MIX keyword.
Note In the case of SCTKW, THERMX, CHEUHA and MIXING the component and
variable indices are redundant as the variable applies to all components.
In the case of SCTALF and SCTMW, the I1 and K1 indices refer to the sample
rather than component range.
For the CHARMF two other sets of indices are needed to indicate the range of
samples over which the particular variable applies. This follows the syntax of
the BIC keyword. For example, CHARMF 1 2 4 5 contains the mole fractions
for components 4 and 5 for samples 1 and 2.
The quantities LL and UL specify limits imposed on the variables during the
regression. All variable limits are applied as a percentage of the current value, scaled
to unity. Thus a maximum value of 1.5 corresponds to allowing a 50% increase in any
variable; a negative value implies that the variable is allowed to change sign (restricted
set!).
Defaults are as follows:
Table 7.8 Default limits for variables

pc 0.5 1.5
Tc 0.5 1.5
Vc 0.5 1.5
Zc 0.5 1.5
Ωa 0.5 1.5
Ωb 0.5 1.5
286 Keywords
VAR
Table 7.8 Default limits for variables (Continued)

ω 0.5 2.0
BIC -5.0 +5.0

si -5.0 5.0
SCTALF 0.5 5.0

SCTMW 0.9 1.1
SCTKW 0.75 1.25
CHEUHA 0.5 1.25
THERMX -5.0 5.0
CHARMW 0.9 1.1
CHARMF .75 1.25
MIXING .75 1.25
Grouped regression variables can be defined by giving them the same variable index.
Examples
Example 1
Regression variables are the critical temperature for component 1, the Ω a values for
components 2 and 3, the binary interaction coefficient for components 1 and 10, and the
volume shift parameters for components 4 and 6 to 10 as a first grouped variable, and
components 1 to 3 and 5 as a second grouped variable. Default limits are taken for all
components except the binary, which is restricted to a lower limit of 30% and an upper
limit of 200% of the current value:
VAR
TCRIT 1 1 1 /
OMEGAA 2 3 1 /
BIC 1 1 10 10 1 0.30 2.00 /
SSHIFT 1 3 1 /
SSHIFT 4 4 2 /
SSHIFT 5 5 1 /
SSHIFT 6 10 2 /
/
That is a total of five variables.
Keywords 287
VAR
Example 2
The default settings for the first five special regression variables with a system of three
samples is as follows:
VAR
’SCTALF’ 1 1 1 0.500000 5.000000 /
’SCTALF’ 2 2 2 0.500000 5.000000 /
’SCTALF’ 3 3 3 0.500000 5.000000 /
’SCTMW’ 1 1 1 0.900000 1.100000 /
’SCTMW’ 2 2 2 0.900000 1.100000 /
’SCTMW’ 3 3 3 0.900000 1.100000 /
’SCTKW’ 0.750000 1.250000 /
’CHEUHA’ 0.500000 1.500000 /
’THERMX’ -5.000000 5.000000 /
/
Note The special regression variables SCTALF and SCTMW cannot be grouped into
one variable and must be entered in the manner shown above, that is one
variable for each sample.
288 Keywords
VAR
VCRIT Specify volumes
Specifies the critical volumes for the components in the system. The keyword is
RUNSPEC
SPLIT RUNSPEC section, and terminated with a forward slash (/) character.
GROUP
COMB The critical volumes are in volume per mole at critical pressure and temperature.
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
OUTECL3
VFP
3 cub ft/lb mole (FIELD),

• UNITS: m /kg-mole (METRIC),
cc/gm mole (LAB), 3
m /kg-mole (PVT-M).
VCRIT values may be entered as an alternative to ZCRIT values - that is one of VCRIT
or ZCRIT should be entered. If VCRIT values are entered, critical Z -factors are
obtained internally using the usual relationship:
p crit V crit
Z crit = --------------------- , [EQ 7.1]
RT crit
where R is the gas constant.
Example
For a 20-component system, in field units:
VCRIT
1.5057 1.4417 1.5698 2.3707 3.2037
4.0847 4.9337 4.9817 5.8948 6.0710
7.8058 8.8374 9.8465 10.830 11.799
13.779 4.759 15.684 16.018 16.018
/
Keywords 289
VCRIT
VCRITVIS Specify volumes for LBC viscosity
calculations
Specifies the critical volumes for the components in the system, to be used in the LBC
RUNSPEC
X SYSTEM viscosity calculation only. The keyword is followed by N c values, where N c is the
SPLIT number of components specified in the RUNSPEC section, and terminated with a
GROUP
COMB forward slash (/) character.
SIMULATE
REGRESS The critical volumes are in volume per mole at critical pressure and temperature.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
3 cub ft/lb mole (FIELD),

• UNITS: m /kg-mole (METRIC),
cc/gm mole (LAB), 3
m /kg-mole (PVT-M).
VCRITVIS values may be entered as an alternative to ZCRITVIS values; that is, one of
VCRITVIS or ZCRITVIS should be entered. If VCRITVIS values are entered, critical
Z -factors are obtained internally using the usual relationship:
p crit V critv
-,
Z critv = ----------------------- [EQ 7.2]
RT crit
Note If VCRITVIS or ZCRITVIS is not entered then values entered with VCRIT or
ZCRIT are used in the viscosity correlation. The form of VCRITVIS is the same
as that of VCRIT.
Example
For a 20-component system, in Field units:
VCRITVIS
1.5057 1.4417 1.5698 2.3707 3.2037
4.0847 4.9337 4.9817 5.8948 6.0710
7.8058 8.8374 9.8465 10.830 11.799
13.779 14.759 15.684 16.018 16.018
/
290 Keywords
VCRITVIS
VERSION Version of PVTi
This keyword indicates the version of PVTi that created the PVI file.
x RUNSPEC
SYSTEM
SPLIT Caution Files in FIELD units containing Differential Liberation (DL) experiments
GROUP
COMB that have Gas Formation Volume Factor (GFVF) observations must be
SIMULATE updated to the current version. See "Converting old projects to the current
REGRESS version" on page 97 for a tutorial on converting old-version PVI files.
BLACKOIL
PSEUCOMP
OUTECL3
VFP
VERSION
2001A /
Keywords 291
VERSION
VFP Start of the VFP section
This is a delimiter keyword, specifying the start of the VFP section.
RUNSPEC
SYSTEM
SPLIT Note This section must follow the SYSTEM section.
GROUP
COMB
SIMULATE
REGRESS
This section is used to generate blackoil tables for VFPi, for any of the currently defined
BLACKOIL fluid samples, by simulating phase and volumetric changes in the wellbore with a
PSEUCOMP constant composition expansion and separator flash at high (reservoir) and low
OUTECL3
X VFP
temperatures. Tables are generated that can then be input into VFPi.
292 Keywords
VFP
WAT100 Output water properties
Specifies the following for calculation of water properties within a black oil table
RUNSPEC
SYSTEM generation for ECLIPSE Black Oil:
SPLIT
GROUP
COMB 1 Temperature at which water properties are required
SIMULATE
REGRESS 2 Pressure at which water properties are required
X BLACKOIL
PSEUCOMP 3 Flag for salt dissolved in water (Y for Yes, N for No)
OUTECL3
VFP 4 Flag for keyword required (S for PVTW, M for PVTWSALT)
5 Flag for units of salinity (P for Parts per million (ppm), G for grammes per liter
(gpl))
6 Flag for gas dissolved (Y for gas, N for no gas)
7 Salinities. This can have just one value if the PVTW keyword is required, but up to
ten values if PVTWSALT is required (zero salinities can be defaulted with *). Units
are given by the above flag.
Examples
Example 1
For PVTW output, salt present of salinity 10000 ppm, with gas dissolved. Temperature
200 °F , pressure 5514.7 psi.
--Water properties requested for E100

WAT100
200.00000 5514.70000 Y S P Y 10000.0000
/
Example 2
For PVTWSALT output, salt present with salinities 10, 20 and 30 gpl, with no gas
dissolved. Temperature 200 °F , pressure 5514.7 psi.
-- Water properties requested for E100

WAT100
200.00000 5514.70000 Y M G N
10.0000 20.0000 30.0000 7*
/
Keywords 293
WAT100
Specifies the following for calculation of water properties within a black oil table
RUNSPEC
SYSTEM generation for ECLIPSE.
SPLIT
GROUP
SIMULATE
BLACKOIL
X PSEUCOMP 3 Flag for salt dissolved in water (Y for Yes, N for No)
OUTECL3
VFP 4 Flag for keyword required (S for PVTW, M for PVTWSALT)
5 Flag for units of salinity (P for Parts per million (ppm), G for grammes per liter
(gpl))
7 Salinities. This can have just one value if the PVTW keyword is required, but up to
10 values if PVTWSALT is required (zero salinities can be defaulted with *). Units
are given by the above flag.
Examples
Example 1

WAT100
200.00000 5514.70000 Y S P Y 10000.0000
/
Example 2
For PVTWSALT output, salt present with salinities 10, 20 and 30 gpl, with no gas
dissolved. Temperature 200 °F , pressure 5514.7 psi.
-- Water properties requested for E200

WAT200
200.00000 5514.70000 Y M G N
10.0000 20.0000 30.0000 7*
/
294 Keywords
WAT200
Specifies the following for calculation of water properties within a run involving EoS
RUNSPEC
SYSTEM output for ECLIPSE Compositional.
SPLIT
GROUP
SIMULATE
BLACKOIL
PSEUCOMP 3 Flag for salt dissolved in water (Y for Yes, N for No)
X OUTECL3
VFP 4 Flag for units of salinity (P for Parts per million (ppm), G for grammes per liter
(gpl))
6 Salinity, units of which are given by the above flag.
Example

WAT300
200.00000 5514.70000 Y P Y 10000.0000
/
Keywords 295
WAT300
WATVFP Output water properties
Specifies the following for calculation of water properties within a run involving black
RUNSPEC
SYSTEM oil output for the VFPi program.
SPLIT
GROUP
COMB 1 Highest temperature at which water properties are required
SIMULATE
REGRESS 2 Highest pressure at which water properties are required
BLACKOIL
PSEUCOMP 3 Lowest temperature at which water properties are required
OUTECL3
X VFP 4 Lowest pressure at which water properties are required
5 Flag for salt dissolved in water (Y for Yes, N for No)
6 Flag for units of salinity (P for Parts per million, G for grammes per liter)
8 Salinity, units of which are given by the above flag.
Example
For PVTW output, salt present of salinity 10000 poem, with gas dissolved. Calculated at
temperatures 200 °F and 80 °F , pressure 6000.0 psi.
--Water properties requested for VFP

WATVFP
200.00000 6000.00000 80.00000 6000.00000 Y P Y 10000.0000
/
296 Keywords
WATVFP
WETGAS Wet gas tables
This keyword requests that ECLIPSE BlackOil wet gas tables be output.
RUNSPEC
SYSTEM The keyword WETGAS has no arguments
SPLIT
GROUP
COMB
SIMULATE
REGRESS
X BLACKOIL
PSEUCOMP
OUTECL3
VFP
X APITRACK
Keywords 297
DEADGAS
WHIT Defines Whitson splitting
This keyword defines data specific to the Whitson splitting algorithm. The method
RUNSPEC
SYSTEM consists of taking the single plus fraction consisting of SCN groups N+ with
X SPLIT composition, mole weight and specific gravity z N+ , M N+ , γ N+ and splitting from SCN
GROUP
COMB
SIMULATE
group N to some large carbon number, say 45, and then re-grouping to a user specified
REGRESS number of pseudo-components N grp .
BLACKOIL
PSEUCOMP
OUTECL3 The keyword is followed by an integer to specify the first SCN group in the plus
VFP fraction, three real numbers for the composition, mole weight and specific gravity and
an integer to specify the required number of pseudo-components on re-grouping, all
terminated with a forward slash (/) character.
Example
Split C 7+ with z 7+ = 0.1 , M 7+ = 140.0 , γ 7+ = 0.85 into three pseudo-components:
WHIT
7 0.10 140.0 0.85 3 /
298 Keywords
WHIT
WHITSON Blackoil tables
Specifies that the Whitson and Torp method for the generation of blackoil tables is used
RUNSPEC
SYSTEM in preference to the Coats method. By default, the program uses the Whitson and Torp
SPLIT method.
GROUP
X COMB The keyword WHITSON has no arguments.
SIMULATE
REGRESS
X BLACKOIL
X PSEUCOMP
OUTECL3
X VFP
Keywords 299
WHITSON
XMFVP/YMFVP
XMFVP and YMFVP ECLIPSE tables
Requests that XMFVP and YMFVP tables, consisting of the liquid and vapor mole
RUNSPEC
SYSTEM
fractions of each component at a set of pressures, be output to the PVO file. The XMFVP
SPLIT and YMFVP keywords have no arguments.
GROUP
COMB
SIMULATE Note You must also define a Constant Volume Depletion (CVD) experiment via the
REGRESS EXP keyword and associated data in order that the XMFVP and YMFVP tables
BLACKOIL
PSEUCOMP can be generated.
X OUTECL3
VFP Only two pressures need be defined in the CVD, which should be the
maximum and minimum required.
300 Keywords
XMFVP/YMFVP
ZCRIT Specify critical Z-factors
Specifies the critical Z -factors for the components in the system. The keyword is
RUNSPEC
SPLIT
GROUP RUNSPEC section, and terminated with a / character.
COMB
SIMULATE
ZCRIT values may be entered as an alternative to VCRIT values, but one of VCRIT or
REGRESS ZCRIT should be entered. If ZCRIT values are entered, critical volumes are obtained
BLACKOIL internally using the usual relationship:
PSEUCOMP
OUTECL3 RT crit Z crit
VFP V crit = ------------------------- , [EQ 7.3]
p crit
Example
For a CO2/Methane/Ethane/I-C4 system:
ZCRIT
0.2709 0.2809 0.2808 0.279 /
Keywords 301
ZCRIT
ZCRITVIS Specify critical Z-factors for LBC
calculations
Specifies the critical Z -factors for the components in the system to be used in the LBC
RUNSPEC
X SYSTEM viscosity correlation only. The keyword is followed by N c values, where N c is the
SPLIT
GROUP number of components specified in the RUNSPEC section, and terminated with a
COMB forward slash (/) character.
SIMULATE
REGRESS ZCRITVIS values may be entered as an alternative to VCRITVIS values, but one of
BLACKOIL VCRITVIS or ZCRITVIS should be entered. If ZCRITVIS values are entered, critical
PSEUCOMP
volumes are obtained internally using the usual relationship:
OUTECL3
VFP
RT crit Z critv
-,
V critv = --------------------------- [EQ 7.4]
p crit

If VCRITVIS or ZCRITVIS is not entered then the values entered with VCRIT or
ZCRIT are used. The form of ZCRITVIS is the same as that of ZCRIT.
Example
For a CO2/Methane/Ethane/I-C4 system:
ZCRIT
0.2709 0.2809 0.2808 0.279 /
302 Keywords
ZCRITVIS
ZI Specify sample composition
Specifies the mole fractions for the components in the system. The keyword is followed
RUNSPEC
X SYSTEM by N c values, and terminated with a forward slash (/) character, where N c is the
SPLIT number of components specified in the RUNSPEC section.
GROUP
COMB
SIMULATE
REGRESS
BLACKOIL
PSEUCOMP
Example
OUTECL3
VFP For a two-component CO2/I-C4 system, 60/40 mixture:
ZI
0.6 0.4 /
Keywords 303
ZI
ZMFVD Composition versus depth table
Requests that a ZMFVD table for ECLIPSE Compositional input, consisting of the
RUNSPEC
SYSTEM composition of each component at a set of depths, be output to the PVO file. The ZMFVD
SPLIT keyword takes no arguments.
GROUP
COMB
SIMULATE Note You must also define a COMPG experiment using the EXP keyword and
REGRESS associated data in order that the ZMFVD table can be generated.
BLACKOIL
PSEUCOMP
X OUTECL3 Only two depths need be defined in the experiment, which should be the
VFP maximum and minimum required.
304 Keywords
ZMFVD
Technical Description
Chapter 8
Overview
This section of the manual contains information on the following:
• "Theoretical background of PVT" on page 306.
• "Analysis techniques" on page 392.
• "Regression in PVT analysis" on page 408
• "Wax and asphaltene precipitation in PVTi" on page 416
• "Cleaning samples contaminated with oil-based mud" on page 420
• "Mixing and recombination of samples" on page 422
Technical Description 305

Overview
Theoretical background of PVT
Introduction
This section of the manual contains information on the following:
• "General background" on page 306.
• "Properties of pure components and mixtures" on page 307.
• "Characterization, splitting and grouping" on page 312.
• "Material balance" on page 330.
• "Flash calculations" on page 330.
• "Equation of state formulation" on page 331.
• "Multiphase flash" on page 338.
• "Viscosity correlations" on page 345.
• "Binary interaction coefficients" on page 353.
• "Blackoil systems" on page 355.
• "Gas condensate systems" on page 358.
• "Process simulation" on page 360.
• "Gas injection processes" on page 361.
• "Variation of composition with depth" on page 363.
• "Practical considerations" on page 365.
• "Theoretical model" on page 369.
• "Blackoil model" on page 372.
• "Pseudo-compositional tables for ECLIPSE GI option" on page 380.
• "Compositional data for ECLIPSE Compositional" on page 381.
• "Water properties" on page 382.
• "Model for API Tracking option in ECLIPSE BlackOil" on page 383.
• "Compositional Data for ECLIPSE Thermal" on page 387.
General background
Both in the fully compositional and extended black oil formulations of reservoir
simulation, an accurate description of the hydrocarbon system and its properties is
important.
In the compositional case the phase behavior of the hydrocarbon system is an integral
part of the simulation. Equal phase fugacity conditions or flash calculations are used
to determine the phase split and composition.
306 Technical Description

In the black-oil and extended black-oil formulations averaged properties of the oil and
gas phases such as B o (oil volume factor), R s (solution gas-oil ratio), B g (gas volume
factor) and R v (condensate gas ratio) are obtained through laboratory experiments or
by using an equation of state description of the hydrocarbon system.
Laboratory tests
Information on oil and gas properties is normally obtained through laboratory tests
upon oil and gas samples. These yield:
1 The mole fraction distribution of lighter components
2 The mole fraction, molecular weight and specific gravity of the residual
3 Pressure-volume relationships obtained through depletion and expansion
experiments.
The mole distribution of the lighter components is usually obtained through gas
chromatography, providing a quantitative separation of the lighter components into
pure compounds with known properties. The heavier components are then lumped
into pseudo-components, for example a C7+ fraction containing hydrocarbons with
carbon numbers from seven upwards.
Beside this simple laboratory analysis more complex methods for analyzing oil and gas
samples are coming into use. In these methods the residual (C7+ fraction) is split up
into sub-fractions through a distillation of the residual. The boiling point of each of
these sub-fractions is then used to assign pseudo component properties through
empirical correlations.
Expansion experiments are often carried out on oil and gas samples from the reservoir
to evaluate volume factors in conditions that reflect the depletion process in the
reservoir as closely as possible.
On the basis of these experimental data, a model composition of the hydrocarbon
mixture can be developed and used in combination with an equation of state model to
calculate additional oil and gas properties at reservoir conditions and at surface
conditions.
The oil and gas volume factors used in black oil type models depend on the process
facilities and configuration. The equation of state model allows black oil tables to be
evaluated for various process facilities and configurations.
An excellent summary of many of the features required to build up an understanding
of fluid phase behavior from sampling considerations through laboratory techniques
to the use of Equations of State can be found in the text of Pedersen et al., and McCain,
[Ref. 33].
Properties of pure components and mixtures

All petroleum accumulations have hydrocarbons as their predominant constituents.
The chemistry of the carbon atom is the richest of all the known elements making it
difficult to assign simple rules to the volumetric behavior of petroleum mixtures. In
order to understand the behavior of mixtures, let us firstly consider the pure
components making up a typical petroleum.
This section contains information on:

• "Pure components" on page 308.
• "Pure component properties" on page 310.
• "Multi-component mixtures" on page 311.
Pure components
The natural gas and crude oil found in underground natural deposits are mixtures of
a large number of different hydrocarbon components and some additional non-
hydrocarbons such as CO 2 , H 2 S , N 2 , H 2 and CO .
Hydrocarbon groups
Hydrocarbons are usually divided into groups based on characteristics in molecular
structure. The most important of these groups are:
Alkanes or paraffins
Alkanes or paraffins are very stable chemical compounds. The chemical formula for
compounds in the group is C n H 2n + 2 . The group contains hydrocarbon components
such as:
Table 8.1 Alkanes
Formula Name
C1H4 Methane
C2H6 Ethane
C3H8 Propane
C 4 H 10 Butane
C 5 H 12 Pentane
C 6 H 14 Hexane
C 7 H 16 Heptane
C 10 H 22 Decane
Each of these molecules can exist as a single straight chain of carbon atoms attached to
three (end carbons), or two (non-end carbons) hydrogen atoms (except Methane,
which has four attached hydrogens).
Isomers
For molecules containing four carbons or more, additional structures other than the
straight chains are possible. The number of these branched molecules or isomers
increases with increasing carbon number.
• Butane has one isomer in addition to the straight-chained normal-Butane, the
branched molecule iso-Butane.

• Pentane has two isomers in addition to the normal-Pentane, known as iso- and
neo-Pentane.
The neo- isomer (cross-shaped) is rarely found in naturally-occurring petroleum.
Most fluid analyses include the iso- and normal- components of Butane and
Pentane.
• Hexane (six carbon atoms) has four isomers in addition to the normal-Hexane.
Note The number of isomers increases rapidly with carbon number making the
identification of the increasingly small concentration of these isomers
impossible.
Napthenes or cycloparaffins
Napthenes are characterized by the chemical formula C n H 2n . Important members in
this family are:
Table 8.2 Napthenes
Formula Name
C3 H6 Cyclopropane
C 5 H 10 Cyclopentane
They share the same formula as the alkenes, which are the same as the alkanes except
that one or more pairs of carbon atoms are linked by double bonds with the consequent
loss of one or more hydrogen atoms.
Note Alkenes and their triple bond equivalents, the alkynes, are rarely found in
naturally occurring petroleum.
Any one of the hydrogens can be replaced by a methyl-( CH 3 ) or higher group, for
example C 7 H 14 , Methylcyclohexane.
Aromatics
The third family of hydrocarbons are aromatics. These are ring-type structures. Some
of the best known members of this family are:
Table 8.3 Aromatics
Formula Name
C6 H6 Benzene
C7 H8 Toluene

SCN groups
All of napthenes, aromatics, more complicated hydrocarbons (polynuclear molecules),
hydrocarbons compounds containing other species, namely Nitrogen, Sulphur,
Oxygen and certain trace metals, are all capable of existing as isomers. Identification of
even a relatively small number of the possible isomers within a petroleum mixture is a
complicated and therefore expensive task.
It is standard practice within the petroleum industry to lump all isomers together on
the basis of the boiling point of the molecule.
Single Carbon Number (SCN) groups are defined for components, for example
hexanes, heptanes, etc. Components are pure hydrocarbon components with normal
boiling point temperatures, that is at one atmosphere pressure, between consecutive
normal-paraffin boiling points.
For example, the SCN hexanes group consists of those hydrocarbons that boil between
the normal boiling points of n-pentane and n-hexane.
Pure component properties

Each of the pure components, which is now taken to include SCN groups for hexane
and above, in a hydrocarbon mixture is characterized by specific physical properties
such as:
Table 8.4 Physical properties
Term Nomenclature
Tc Critical temperature
pc Critical pressure
Vc Critical volume
Zc Critical Z -factor
Mw Molecular weight
ω Acentric factor
Tb Normal boiling point
ρ ref Reference density (usually specified at a reference temperature)
T ref Reference temperature
[P] Parachor (see surface tensions)
The component library in PVTi contains properties for some of the more common pure
components.

Multi-component mixtures
As mentioned previously, natural gas and crude oil contains literally thousands of
different components. No attempt is usually made to identify all the hydrocarbons
beyond C 5 : rather, SCN groups denoted hexanes, heptanes, etc., are used. Even so, you
may be left with ten or more components in any given petroleum mixture.
To describe a multi-component system thermodynamically using an equation of state
model, the system must be defined in terms of the properties of the components and
their mole fractions in the mixture.
The terms and nomenclature used to characterize multi-component mixtures may be
th
summarized as follows, where the subscript i denotes the i component:
Table 8.5 Multi-component (ii) mixtures
Term Nomenclature
n Number of components in the mixture
T c, i Critical temperature
p c, i Critical pressure
Table 8.6 Multi-component (ii) mixtures
Term Nomenclature
M w, i Molecular weight
ωi Acentric factor
xi Mole fraction in the liquid phase
yi Mole fraction in the vapor phase
zi Mole fraction in the mixture as a whole
p Pressure of the mixture
T Temperature of the mixture
V Mole fraction of the vapor phase
L Mole fraction of the liquid phase
Note Not all of the parameters listed above are independent.
The thermodynamic behavior of a multi-component mixture depends strongly on

composition, pressure and temperature.

Phase diagrams
In terms of a phase diagram plotted against pressure and temperature, the upper part
of the phase envelope, up to the critical point, represents the bubble-point curve.
From the critical point the phase envelope continues as dew point curve. The phase
envelope encircles the two phase region in the phase diagram.
A phase diagram is often characterized by the maximum pressure, the cricondenbar,
and the maximum temperature, cricondentherm.
The area within the phase envelope bounded by a vertical line through the critical
point and the line represented by the cricondentherm is the retrograde area. A
reduction of pressure into this area results in retrograde condensation.
For the RK and ZJ equations of state the critical point calculation implements the
theory developed in [Ref. 51].
For the PR and SRK equations of state the calculations use the theory developed in the
follow-up paper in [Ref. 48].
Characterization, splitting and grouping

This section of the manual contains information on:
• "Components" on page 312.
• "Characterization" on page 313.
• "Splitting" on page 313.
Components
All components in PVTi are labeled as being one of three possible types:
• LIB: Library
• CHAR: Characterization
• USER: User defined
LIB components
Lib components are the commonly occurring hydrocarbons C 1 , C 2 ,..., C 45 and the
specific non-hydrocarbons H 2 O , H 2 S , CO 2 , N 2 , H 2 and CO .
CHAR components
Components designated as CHAR are typically the last component in a PVT report and
as such comprise the plus fraction, that is, contains all remaining hydrocarbons. For
such fractions, a laboratory would typically measure only the mole weight and specific
gravity, denoted here by M N+ , γ N+ , where N is the number of the first carbon group
in the fraction.
User components
User components are those for which you must supply all the data necessary to define
the equation of state parameters, that is T c , p c , V c , M w and ω .

Characterization
Clearly, correlations are required that are capable of generating the critical properties
and acentric factors from just the mole weight and specific gravity.
In general, any p - V - T report gives at least the mole weight of the plus fraction. If this
is the only information available, PVTi estimates a value of the specific gravity from
known (SCN) distributions. Otherwise, any two out of the four of M w , γ , T b , K w is
sufficient to characterize a component.
PVTi employs five sets of correlations for generating critical properties by:
• Kesler and Lee [Ref. 10]
• Cavett [Ref. 11]
• Riazi and Daubert [Ref. 12]
• Winn [Ref. 43] and [Ref. 44]
• Pedersen [Ref. 42].
It also uses four sets of correlations for acentric factors by:
• Lee and Kesler [Ref. 13]
• Edmister [Ref. 14]
• Thomassen [Ref. 30]
• Pedersen [Ref. 42].
Watson characterization factor

In many of these correlations, a quantity often quoted is the Watson characterization
factor, denoted K w which is defined as:
3
T bi = ( K w γ i ) [EQ 8.1]
where:
• T bi is the normal boiling point temperature (in degrees Rankine)
th
• γ i is the specific gravity of the i component.
Splitting
The plus fraction often has an importance that appears to far outweigh its relatively
small mole fraction of a fluid sample. In particular, saturation pressure calculations can
be extremely sensitive to the mole fraction and properties of the plus fraction. More
accurate predictions requiring less regression of equation of state parameters can be
achieved if a thorough description of the plus fraction can be made.
Ideally, a complete true-boiling-point distillation (TBP) should be made that yields not
only the detailed composition of the plus fraction but also the boiling points, specific
gravities and molecular weights of the constituent components. However, this analysis
is rarely performed and so a general procedure to describe the distribution of
components and properties within a plus fraction is required.

Three techniques are available within PVTi for the splitting of the plus fraction into
sub-fractions:
• Constant Mole Fraction (CMF);
• Whitson.
• Semi-Continuous Thermodynamics (Modified Whitson)
All of these techniques rely on a probability density function (PDF) to relate mole
fraction to mole weight. This approach is due to Whitson [Ref. 4].
Probability density model

The PDF used by Whitson to describe the relation between mole fraction and mole
weight is a three-parameter gamma function:
(α – 1) (M – η)
(M – η) exp -------------------
β
p ( M ) = ---------------------------------------------------------------------- [EQ 8.2]
α
β Γ(α)
where:
• α , β , η are parameters defining the distribution
• Γ is the gamma-function
• α gives a measure of the shape of the distribution
• η is the lowest mole weight in the plus fraction
• β is a normalization condition that can be determined from the condition:
M N + – η = αβ [EQ 8.3]
where M N + is the average mole weight of the plus fraction.
The cumulative probability function, P ( X ≤ x ) is the integral of p ( x ) from η to x :

x
P( X ≤ x) = ∫ η p ( x ) dx [EQ 8.4]
that is evaluated numerically from:

∞ α+j
–y y
P( X ≤ x) = e ∑ ------------------------------
Γ(α + j + 1)
[EQ 8.5]
j=0
where y = ( x – η ) ⁄ β
The frequency, f i of a component i having mole weight boundaries M i – 1 , M i is given

by the integral:
Mi
fi = ∫M p ( M ) dM [EQ 8.6]
i–1
= P ( M ≤ Mi ) – P ( M ≤ Mi – 1 ) [EQ 8.7]
and the mole fraction z i is related to its frequency by:

zi = zN + fi [EQ 8.8]
The average mole weight in the same interval is given by:

P ( M ≤ M i ,α + 1 ) – P ( M ≤ M i – 1 ,α + 1 )
M i = η + αβ -------------------------------------------------------------------------------------------------- [EQ 8.9]
P ( M ≤ M i ,α ) – P ( M ≤ M i – 1 ,α )
where the P ( X ≤ x ) functions all use the same value of β regardless of the value of α ,
(that is α or α + 1 ) and where z N + is the total mole fraction of the plus fraction.
In the absence of any other data, you should assume that α is unity. The parameter η
is the minimum mole weight that occurs in the plus fraction, therefore if the plus
fraction were C 7+ , then a good estimate to η is 92.0.
Generally, Whitson recommends:

η = 14n – 6 [EQ 8.10]
where n is the first SCN group in the plus fraction, for example 7.
Note Both splitting techniques available in PVTi use the Whitson PDF to define the
mole fraction/mole weight distribution of the plus fraction, however, they
differ slightly in the way they sub-divide into the required number of pseudo-
components.
Constant mole fraction, CMF

In this technique, you select the number of pseudo-components required from the split.
From this PVTi calculates the mole fraction to be assigned to each of the pseudo-
components, scaled to unity.
Note By default equal mole fractions are assigned.
Generally, you may consider skewing the distribution thus: highest mole fraction for
the lightest pseudo-component to smallest mole fraction for the heaviest pseudo-
component.
Hint If you have evidence on the shape of the plus fraction distribution, it may be
prudent to manually change the skewness parameter α , that is less than unity
for a condensate and greater than unity for a crude oil.
The value of the minimum mole weight in the plus fraction η rarely needs changing
from its (calculated) default value.
Knowing the required mole fraction split, PVTi then integrates the PDF from η up to
some mole weight M 1 such that integral gives the correct first mole fraction, and then
similarly for the remaining pseudo-components.
Knowing the mole weight boundaries, the program can calculate the average mole
weights of the pseudo-components, estimate the specific gravities from the SCN
distributions and calculate the critical properties, etc., from the various correlations
currently in force.

Whitson
The Whitson technique consists of splitting the plus fraction into SCN groups N (the
first in the plus fraction) to some high number, for example 45.
The mole weight boundaries are estimated from the user specified plus fraction mole
weight and specific gravity from which one can calculate a plus fraction Watson factor:
0.15178 – 0.84573
K w, N+ = 4.5579M w, N+ γ N+ [EQ 8.11]
Assuming that an average Watson factor K w, N+ can be applied to each individual SCN
groups, the SCN groups’ specific gravities can be estimated from equation (1), namely:
1⁄3
T b, i
γ i = ----------------
- [EQ 8.12]
K w, N+
where the T b, i are the normal boiling point temperatures of the SCN groups, which in
fact define the SCN groups. That is SCN C 6 is all hydrocarbons that have normal
boiling point temperatures between those of normal- C 5 and normal- C 6 , etc.
Then, with known boiling points and specific gravities, mole weights can be estimated
using the Riazi-Daubert correlation, which is of the general form:
b c
θ = aT b γ [EQ 8.13]
where θ is the property to be evaluated, say M w and ( a ,b ,c ) are tabulated coefficients,

see [Ref. 4].
With average mole weights for the SCN groups, the boundary mole weights are
estimated from:
1
M i = --- ( M i – 1 + M i ) [EQ 8.14]
2
These can then be used to integrate the PDF to give mole fractions for each of the SCN
groups.
Hint Generally, you will not want to work with 20-30 additional components.
Therefore the SCN groups are pseudoised down into Multi-Carbon Number
(MCN) groups. A method for estimating the required number of MCN groups
is that due to Sturge’s, discussed in [Ref. 4]. However, this often predicts the
use of four or five pseudo-components. It has been our experience that two or
three pseudo-components is sufficient for most purposes.
Multi-carbon number (MCN) groups

In evaluating the properties of the MCN groups from the basis SCN groups, simple
mole weighted averages are taken, that is.:
i = lN
1
M l = ----
zl ∑ zi Mi [EQ 8.15]
i = l1

where the summation is performed over the SCN groups i in the MCN group l and z l
th
is the mole fraction of the l MCN group.
Modified Whitson splitting (1988)

The Constant Mole Fraction (CMF) and Whitson Splitting methods described above
both consist of a continuous molar distribution model, which is subsequently
discretized into a set of pseudo-components.
Another method to effect the discretization is to use quadrature methods; in particular,
integrals of the form:
N
∞ –x
∫0 f ( x )e dx = ∑ wif ( xi ) , [EQ 8.16]
i=1
where the weighting factors w i and quadrature points x i are determined from a class
of Laguerre polynomials for a given order N , see [Ref. 31].
For our mole weight/mole composition model, we can associate the quadrature points
x i with pseudo-component mole weights M i = η + βx i and the mole fractions
zi = wi f ( xi ) .
To account for compositional variations with depth, Whitson extended the standard
model to account for these variations, or plus fraction mole weight and skewness
parameter. Since different plus fraction mole weights and skewness parameters lead to
different β parameters, the following modification was introduced:
β0
β = --------------------------
- [EQ 8.17]
[ 1 + ln ( δ ) ]
where β 0 and δ are parameters in the modified distribution function p 0 ( M ) given by:
(α – 1) α
(M – η) – ( M – η ) ( 1 + ln ( δ ) )
p 0 ( M ) = ----------------------------------- exp ---------------------- ---------------------------------- = p ( M ) [EQ 8.18]
α β0 [ ( M – η ) ⁄ β0 ]
β0 Γ ( α ) δ
that is numerically identical to the original function p ( M ) , see[EQ 8.2].

To determine the mole fraction, we integrate the PDF by Gaussian quadrature by
making the following transformation:
M–η
x 0 = -------------- [EQ 8.19]
β0
with dM = β 0 dx to give:
( α – 1 ) –x0
∞ x0 e
∫0 ----------------------------------------------------
x –α 0
dx = 1 [EQ 8.20]
Γ ( α )δ 0 ( 1 + ln ( δ ) )
Comparing with [EQ 8.16], we have:

(α – 1) α
x0 ( 1 + ln ( δ ) )
f ( x 0 ) = ------------------------------------------------- [EQ 8.21]
x
Γ ( α )δ 0
and thus the mole fraction and mole weight are given by:
z i = w i f ( x 0i ) [EQ 8.22]
M i = η + β 0 x 0i [EQ 8.23]
Procedure
The procedure for using this method is particularly attractive.
Firstly we choose the number of pseudo-components required, N , which is normally
in the range 2 ≤ N ≤ 5 : this then fixes the values of weights and points, w i and x i .
The minimum mole weight η is chosen as before.
Next we specify β 0 which is done by selecting the maximum mole weight component
to be used: Whitson recommends M N = 500.0 but we prefer 2 × M N+ . Either way we
now have:
MN – η
β 0 = -----------------
- [EQ 8.24]
x 0N
Next we estimate a value for the skewness parameter, which as before is taken as unity
in the absence of any other information. Then to satisfy the total number of moles and
mass, we calculate:
β
δ = exp ⎛ -----0- – 1⎞ [EQ 8.25]
⎝β ⎠
and calculate the mole fractions and mole weights. Finally, the average plus fraction
mole weight is checked:
N
∑ zi Mi
=1
M N+ = i---------------------
- [EQ 8.26]
N
∑ zi
i=1
and the δ is adjusted if [EQ 8.26]does not equate to the measured value.
The appeal of this model is it allows a variable plus fractions mole weight and
skewness parameter by sample composition. That is different plus fraction mole
weight and mole fraction, but allows us to use a common set of pseudo-components
with fixed mole weights and hence critical properties.

Hint Having used this procedure to characterize the plus fraction(s), you can then
use the special regression facility, which allows the sample plus fraction mole
weights and skewness parameters to be variables.
In particular, where there is a known variation of composition with depth and

you have at least two samples from different depths in the hydrocarbon
column, this technique can be used with some success, see [Ref. 32].
Special regression facility

An additional special regression facility has been added to this model to allow different
characterizations of the split pseudo-components from this technique.
Having obtained mole fractions and mole weights from the above method, a variant
on the Watson K factor may be derived, denoted F c .
This is then used to estimate specific gravities and hence critical properties, acentric
factors, etc., as per the original Whitson model. The additional feature is the new
characterization factor, F c , which can be regressed, that is:
tot cal reg

Fc = Fc × Fc [EQ 8.27]
cal
where F c is the value calculated by the Whitson model, that is assuming SCN-cuts,
reg
and F c is the potentially regressable function which defaults to unity. This allows
you to make your fluid more or less aromatic with respect to the standard distribution
(equivalent to a K w ∼ 12.0 ), which in turn yields different critical properties, etc., and
hence different fluid behavior.
Grouping
Grouping of components is performed by one of three techniques:
• Molar averaging
• Weight averaging
• Mixing rule.
All of these are explained in the comprehensive text of Joergensen and Stenby [Ref. 45].
Consistency checks and correlations

PVT analysis of reservoir fluids is usually performed so that a model fluid, be it
blackoil, compositional or otherwise, can be constructed for use in a reservoir
simulator.
This analysis requires that an equation of state model is used to match measured data
from laboratory experiments.
Any uncertainties in the laboratory data, brought about by inconsistencies in the
laboratory measurements but more likely due to problems encountered in the taking
of the fluid samples, feed through into a poor fluid model.

Consistency tests
Most fluids that show some form of compositional behavior, namely gas condensates
and volatile oils, are subject to a constant volume depletion experiment (CVD) as part
of their analysis. For further information see "Compositional material balance" on page
320.
A material balance calculation can be performed using the data that is generated from
such an experiment from which quantities such as liquid compositions, K -values,
molecular weights and densities of vapor and liquid, etc., can be evaluated. This data
can then be examined to look for any inconsistencies, such as lack of monotonicity.
Additional data often given in a laboratory PVT report are separator compositions for
the liquid and vapor streams. If such data is available, then K -values can be
constructed and plotted against theoretical fits, again as a test of consistency.
Finally, estimates of recovery of gas and oil from the reservoir can be made as well as
generation of blackoil tables, without recourse to the use of the equation of state model.
Compositional material balance

Several authors have published models for testing the consistency of laboratory CVD
data, [Ref. 19] and [Ref. 6].
• "Consistency checks and correlations" on page 319.
• "Liquid compositions and K-values" on page 321.
• "Physical properties" on page 324.
• "Correlations" on page 324.
• "Recombination of separator data" on page 327.
Note The model employed in PVTi was adapted from a program developed by
Pearce [Ref. 20].
A typical laboratory CVD report might be as follows [Ref. 6]:
Table 8.7 CVD Report
Equil
Equilibrium vapor Pressures (psi)
Liquid
Component 6764.7 5514.7 4314.7 3114.7 2114.7 1214.7 714.7 714.7
Carbon Dioxide 2.37 2.40 2.45 2.50 2.53 2.57 2.60 .59
Nitrogen 0.31 0.32 0.33 0.34 0.34 0.34 0.33 .01
Methane 73.19 75.56 77.89 79.33 79.62 78.9 77.8 12.42
Ethane 7.80 7.83 7.87 7.92 8.04 8.40 8.70 3.36
Propane 3.55 3.47 3.40 3.41 3.53 3.74 3.91 2.92
isoButane .71 .67 .65 .64 .66 .72 .78 .91

Table 8.7 CVD Report (Cont.)
Equil
Equilibrium vapor Pressures (psi)
Liquid
Component 6764.7 5514.7 4314.7 3114.7 2114.7 1214.7 714.7 714.7
n-Butane 1.45 1.37 1.31 1.30 1.33 1.44 1.56 2.09
isoPentane .64 .59 .55 .53 .54 .59 .64 1.4
n-Pentane .68 .62 .58 .56 .57 .61 .66 1.6
Hexanes 1.09 .97 .88 .83 .82 .85 .9 3.68
Heptanes plus 8.21 6.2 4.09 2.64 2.02 1.84 2.12 71.01
Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
Mw(C7+) 184.0 160.0 142.0 127.0 119.0 115.0 114.0 213.0
Sg(C7+) .816 .799 .783 .770 .762 .758 .757 .833
Zvap 1.238 1.089 .972 .913 .914 .937 .960
Nprod 0.000 9.024 21.744 38.674 55.686 72.146 81.301
Sliq 0.00 14.1 19.70 21.60 21.30 20.20 19.3
This gas condensate with an assumed reservoir composition equal to composition at

the first pressure, which corresponds to the saturation pressure of the fluid,
p sat = 6764.7 psia is put through the stages of the CVD.
The CVD consists of reducing the pressure to the required stage pressure, and then
removing gas so that the volume remaining in the cell is equal to volume at the
saturation pressure.
At each stage, the composition, mole weight and specific gravity of the plus fraction
(denoted Mw and Sg), the vapor Z -factor (denoted Zvap), the number of moles
produced (denoted Nprod) and the liquid saturation (denoted Sliq), are measured.
The subsequent analyses used to generate liquid compositions, K -values, volumes,
masses and densities follow those of Whitson and Torp [Ref. 6].
Liquid compositions and K-values

Cell compositions
A material balance performed on the total moles and individual components is:
n tj = n lj + n vj [EQ 8.28]
n tj z ij = n lj x ij + n vj y ij [EQ 8.29]
where:
• n tj , n lj , n vj are the total, liquid and vapor moles at stage j
• ( p = p j ) and z ij , x ij , y ij are the total, liquid and vapor mole fractions of

component i .

Starting with one mole of fluid at the saturation pressure (first stage in the reported
th
CVD), the number of moles remaining the cell after the j stage is:
j
n tj = 1 – ∑ Δn pk [EQ 8.30]
k=2
where n pk is the number of moles of gas produced at stage, denoted Nprod in the table
th
above, so that Δn pk is the incremental gas recovered during the k stage, that is,
reducing the pressure from p k – 1 to p k .
th th
Similarly, the mole fraction of the i component remaining in the cell after the j
stage is:
j
n tj z ij = z i1 – ∑ Δn pk y ik [EQ 8.31]
k=2
Cell volumes
To convert moles and mole fractions to volumes, one must have some reference
volume, such as the volume of the cell.
For a gas condensate, the cell volume, for one mole, can be calculated from the real gas
law:
Z 1 RT
V cell = ------------- [EQ 8.32]
p1
where:
• Z 1 and p 1 are the vapor Z -factor and pressure at the first (saturation) pressure
• R is the universal gas constant
• T is the temperature at which the CVD is performed, assumed constant

throughout the experiment.
For a volatile oil, a CVD report gives the molecular weight and density of the bubble
point liquid, M w1 and ρ 1 , and thus the cell volume, again for one mole, is calculated
as:
M w1
V cell = ----------- [EQ 8.33]
ρ1
th
Knowing the cell volume and the liquid saturation at the j stage, S lj (denoted Sliq in
the table above), the volume of the cell occupied by liquid is:
V lj = S lj V cell [EQ 8.34]
and thus the volume of the cell occupied by gas is:

V vj = ( 1 – S lj )V cell [EQ 8.35]

Moles and mole fractions
th
The number of moles of vapor remaining in the cell at the j stage can be determined
from the real gas law:
p j V vj
n vj = ------------
- [EQ 8.36]
Z j RT
th
Thus, the composition of the liquid remaining in the cell at the j stage can be
determined from:
n tj z ij – n vj y ij
x ij = -------------------------------
- [EQ 8.37]
n tj – n vj
Equilibrium K-values
The K -values can now be estimated from the ratio of the vapor and liquid mole
fractions:
y ij
K ij = -----
- [EQ 8.38]
x ij
The best test of the consistency of the CVD data is to plot the logarithm of the p j K ij
product against the characterization factor F , [Ref. 20], where:
F = b i ⎛ -------- – ---⎞
1 1
[EQ 8.39]
⎝T ⎠
bi T
where T is the reservoir or CVD experiment temperature and b i is the Hoffmann-

th
Crump-Hocott b -factor for the i component, which is given by:
( log ( p ci ) – log ( p ref ) )

b i = ----------------------------------------------------
- [EQ 8.40]
⎛ -------
1
- – -------⎞
1
⎝T ⎠
bi T ci
Here:
• T ci and T bi are the critical and normal boiling point temperatures
• p ci is the critical pressure (in psia)
• p ref is the reference pressure, usually 14.7 psia.
Plots
As pointed out by Bashbush, [Ref. 19], the curves should plot in a parallel-like trend,
that is, there should be no humps in any individual curve or crossing of any pair or
pairs of curves.
The highest curve should belong to nitrogen, followed by methane and carbon dioxide.
Then depending on the fluid composition, either ethane or hydrogen sulphide should
be plotted followed by the curves for the remaining hydrocarbon components in order
of increasing molecular weight.

Finally, it should be noted that the iso- components of butane and pentane should plot
above their corresponding normal components. This plot is available from PVTi.
Other useful plots consists of fingerprint plots of the liquid composition remaining in
the cell at the end of the CVD (the last column in the above table) and the equivalent
calculated liquid composition, and the calculated liquid compositions as a function of
pressure. Both of these plots are available from PVTi.
Physical properties
As well as performing material balances on moles and mole fractions, material balance
can be done on the total, liquid and vapor masses:
m tj = m lj + m vj [EQ 8.41]
th
The total mass remaining the cell at the j stage is given by:
j
m tj = M s – ∑ Δn pk M vk [EQ 8.42]
k=2
where M s and M vk are the initial or saturation pressure molecular weight and the
th
vapor molecular weight at the k stage, respectively, both of which can be evaluated
from Kay’s rule:
Nc
M vj = ∑ y ij M i [EQ 8.43]
i=1
th
where M i is the molecular weight of the i component.
Vapor molecular weight is calculated from:

m vj = n vj M vj [EQ 8.44]
And liquid molecular weight from the mass balance:

m lj = m tj – m vj [EQ 8.45]
Now, knowing the masses and volumes of liquid and vapor in the cell at each stage,
one can calculate densities from ρ = m ⁄ V
Plots of these quantities are available from PVTi.
Correlations
Alani-Kennedy liquid densities
The liquid molar volume of a pure hydrocarbon component is given by the smallest
real positive root of the Van der Waal’s equation:
⎛ a⎞
RT = ⎜ p + ------⎟ ( V – b ) [EQ 8.46]
⎝ V ⎠
2

that contains the two unknown parameters ( a ,b ) .
Alani and Kennedy, [Ref. 20], inserted known values of T , p and V into [Ref. 16] for
two pressures for a variety of hydrocarbons. Thus they were able to observe the
variations of ( a ,b ) from which they derived:
a = K exp ( n ⁄ T ) [EQ 8.47]
b = mT + C [EQ 8.48]
where:
• ( n ,K ,m ,C ) are tabulated constants for the hydrocarbons
• C 1 , C 2 ,..., C 6 : the coefficients for the plus fraction
• C 7+ can be calculated from given correlations in molecular weight and specific

gravity of the plus fraction and temperature (see [Ref. 22]).
The ( a ,b ) coefficients for a hydrocarbon mixture are given by Kay’s rule:
C N+ C N+
a mix = ∑ xi ai b mix = ∑ xi bi [EQ 8.49]
C1 C1
where x i are the liquid mole fractions.
Note The non-hydrocarbons N 2 , CO 2 and H 2 S are added into the C 1 fraction.
Vapor Z-factor
The vapor Z -factor, Z vap , can be estimated from the correlation due to Dranchuk,
Purvis and Robinson [Ref. 21]:
0.27p
Z vap = ---------------r- [EQ 8.50]
ρr Tr
where T r and p r are the reduced temperature ( = T ⁄ T c ) and reduced pressure

( = p ⁄ p c ) , and ρ r is the pseudoreduced density, which is found by iteratively solving:
f ( ρr ) = 0 [EQ 8.51]
where:
6 3 2 3 2 2
f ( ρ r ) = Aρ r + Bρ r + Cρ r + Dρ r + Eρ r ( 1 + Fρ r ) exp ( – F ρ r ) – G [EQ 8.52]
and:
A = 0.06423 [EQ 8.53]
B = 0.5353T r – 0.6123 [EQ 8.54]

0.5783
C = 0.3151T r – 1.0467 – ---------------- [EQ 8.55]
2
Tr
D = Tr [EQ 8.56]
0.6816
E = ---------------- [EQ 8.57]
2
Tr
F = 0.6845 [EQ 8.58]
G = 0.27T r [EQ 8.59]
with the initial estimate for the pseudoreduced density of:
0.27p
ρ r0 = ---------------r- [EQ 8.60]
Tr
Pseudo-critical temperatures and pressures

Pseudo-critical temperatures and pressures for a fluid mixture can be calculated from
one of two methods:
• Kay’s rule
• Wichert and Aziz Correlation.
Kay’s rule has been stated already, being a mole fraction weighted sum of the
appropriate quantity.
The correlations due to Wichert and Aziz, [Ref. 22], correct for the presence of the so-
called sour gases, CO 2 , H 2 S and N 2 .
The pseudo-critical temperature and pressure of the hydrocarbon portion of a

condensate is given by:
HC HC HC 2
Tc = 187 + 330γ g – 71.5 ( γ g ) [EQ 8.61]
HC HC HC 2
pc = 706 + 51.7γ g – 11.1 ( γ g ) [EQ 8.62]
HC
where γ g is the hydrocarbon gas gravity that is related to the mixture gas gravity,
mix
γg by:
mix
γ g – 0.9672y N – 1.5195y CO – 1.1765y H S
HC 2 2 2
γ g = -----------------------------------------------------------------------------------------------------------------
- [EQ 8.63]
1 – y N – y CO – y H S
2 2 2
with:
γg
mix ∑ y i M-i
= ----------------- [EQ 8.64]
M air
and M air is the molecular weight of air (= 28.97).

The pseudo-critical temperature and pressure for the mixture, including non-
hydrocarbons, is:
HC
T pc = ( 1 – y N – y CO – y H S )T c + 227.3y N + 547.6y CO + 672.4y H S [EQ 8.65]
2 2 2 2 2 2
and
HC
p pc = ( 1 – y N – y CO – y H S )p c + 493.0y N + 1071.0y CO + 1306.0y H S [EQ 8.66]
2 2 2 2 2 2
The correction to T pc and p pc in the presence of sour gas ( CO 2 and H 2 S ), due to

Wichert and Aziz, is:
+ 15 ⎛ y H S – y H S⎞
0.9 1.6 0.5 4.0
ε = 120 ( y CO + y H S ) – ( y CO + y H S ) [EQ 8.67]
2 2 2 2 ⎝ 2 2 ⎠
to give:
T∗ pc = T pc – ε [EQ 8.68]
and
p pc ( T pc – ε )
p∗ pc = --------------------------------------------------------
- [EQ 8.69]
T pc + y H S ( 1 – y H S )ε
2 2
Gas viscosity
The dynamic gas viscosity of a hydrocarbon gas at a temperature T can be estimated
from a correlation due to Lee, Gonzalez and Eakin, [Ref. 24]:
C
μ g = A exp ( Bρ ) [EQ 8.70]
where:
–4 1.5
10 ( 9.4 + 0.02M w )T
A = ------------------------------------------------------------
- [EQ 8.71]
( 209.0 + 19.0M w + T )
B = 3.5 + 986.0
------------- + 0.01M w [EQ 8.72]
T
C = 2.4 – 0.2B [EQ 8.73]
and:
pM w
ρ = -----------
- [EQ 8.74]
ZRT
3
where ρ is measured in g/ cm .
Recombination of separator data

In most laboratory PVT reports, the well stream fluid used as the basis for the
experiments is generally a recombined sample of separator gas and oil. The rates of
production of the two phases are also recorded and the ratio of gas rate to oil rate
defines the separator gas-oil ratio (GOR).

Generally, the separator is operating at some pressure and temperature above what
would be considered as standard conditions, such as 60 °F , 14.7 psia, in which case the
separator oil is almost certainly not be stable at standard conditions since it probably
still contains dissolved gas. The best laboratory analyses contain a mole fraction
analysis of the separator oil and gas and the stabilized, standard conditions, oil and
gas.
Any available gas and oil compositional analysis can be used to generate a set of K -
values. These K -values can be plotted using the Hoffmann-Crump-Hocott
construction, see [Ref. 20] and"Correlations" on page 324.
An analysis of sets of K -values from many fluid samples by Standing, [Ref. 25],
indicated that for pressures less than 1000 psia, compositional dependency in K -
values, usually indicated by the apparent convergence pressure of the log ( K ) versus
log ( p ) plot, is small and can often be neglected.
Noting this, Standing fitted the following form:
1 ( a + cF )
K = --- 10 [EQ 8.75]
p
with:
–4 –8 2
a = 1.20 + 4.5 ×10 p + 15.0 ×10 p [EQ 8.76]
–4 –8 2
c = 0.89 – 1.7 ×10 p – 3.5 ×10 p [EQ 8.77]
where F is the Hoffmann et al. characterization factor, refer to [EQ 8.39].

As most separators are operated at pressures less than 1000 psia, the so-called Standing
K -values provide a reasonable test of the quality of the calculated separator data and
can be used as an alternative for blackoil table generation etc., when no such actual
data is available.
The feed to each stage of the separator can be calculated given the compositions of
liquid and vapor resulting from the separator flash, the GOR and the liquid density. It
can be shown that the phase split β is given by F ⁄ ( 1 + F ) where:
( GP std M )
F = ---------------------------
- [EQ 8.78]
( RT std ρ oil )
where:
• G is the GOR (gas at STC, oil at stage conditions)
• M is the molar weight of oil, that is ∑ xi Wi

i
• d ρ oil is the density of the separator oil.
Given β , the feed z i can be found from z i = βy i + ( 1 – β )x i .

For reservoirs whose initial pressure, p init , is greater than its saturation pressure, p sat ,
estimates of fluid recovery can be made if the Z -factors in the range p init ≥ p ≥ p sat are
known.
Note Above the saturation pressure, no compositional changes in the fluid take
place so for a unit volume of the fluid at isothermal conditions:
n1 Z1 n2 Z2
-----------
- = -----------
- [EQ 8.79]
p1 p2
Starting with 100 moles of fluid at the initial reservoir pressure, the number of moles
at some pressure p j above the saturation pressure is:
left rem
nj = 100 – n j [EQ 8.80]
where:
rem ⎛ p j⎞ ⎛ Z 1⎞
nj = 100 ⎜ -----⎟ ⎜ ------⎟ [EQ 8.81]
⎝ Z j⎠ ⎝ p 1⎠
At pressures below the saturation pressure, the produced moles are known from the
CVD experiment.
The fluid at each stage can then be flashed to surface conditions, using separator or
Standing’s K -values to get the volumes of liquid and vapor that could be expected at
the surface, assuming 100% efficiency in the separators.
An option exists to produce liquid along with gas, in a ratio defined by input relative
permeabilities, using the technique described in Reudelhuber and Hinds, refer to [Ref.
37].

Equation of state
General information
PVTi allows you to fit data to an Equation of State. See "The fluid model" on page 109
and "Batch system and keywords" on page 163 for further information on how to
define Equations of State in PVTi.
• "Material balance" on page 330.
• "Flash calculations" on page 330.
• "Equation of state formulation" on page 331.
• "Surface tensions" on page 334.
• "Three-parameter equation of state" on page 335.
• "Binary interaction coefficients" on page 353.
Material balance
We consider a hydrocarbon system consisting of one mole of mixture or feed of
composition z i .
In general, this consists of:

• L moles of liquid of composition x i
• V moles of vapor of composition y i
• at some pressure and temperature, p and T .

The conditions exist that:
L+V = 1 [EQ 8.82]
∑ xi = ∑ yi = ∑ zi = 1 [EQ 8.83]
i i i
Lx i + Vy i = z i [EQ 8.84]
Flash calculations
For a thermodynamic system to be in equilibrium it is further required that the
fugacities in the liquid and vapor phases must be equal for each component:
f iL = f iV [EQ 8.85]
The equations [EQ 8.82] to [EQ 8.85] in the unknowns L , V , x i and y i may be solved
to find the equilibrium state of a system of known composition z i .

Equation of state
The fugacities are functions of temperature, pressure and composition (see "Equation
of state formulation" on page 331),
f i = f i ( T , p, x i)
The fugacities can be calculated directly from an equation of state. Defining

equilibrium constants K i for each component as:
y
K i = ----i
xi
the mole fractions of each component in the liquid and vapor phases are defined as:
zi
x i = ------------------------------------- [EQ 8.86]
[ 1 + ( K i – 1 )V ]
and
Ki zi
y i = ------------------------------------- [EQ 8.87]
[ 1 + ( K i – 1 )V ]
The flash calculation determining the equilibrium conditions in the two-phase region
is performed in two stages:
Use Michelsen’s stability criterion, [Ref. 16], to establish how many phases are present.
The stationary condition used to determine the Gibb’s energy minimum is obtained
using successive substitution accelerated with the general dominant eigenvalue
method, [Ref. 8], switched to BFGS minimization near the minimum, [Ref. 9].
If two phases are present, solve the phase split calculation to obtain equal fugacity
liquid and vapor states. The solution variables used are J i = log ( K i ) and the vapor
mole fraction. The stability check provides an initial set of values, which are refined
using accelerated successive substitutions, and finally solved using a full Newton-
Raphson method.
In the case of finding a saturation pressure, that is a bubble point for an oil or a dew
point for a condensate, the same N -fugacity equations are used, [EQ 8.85], and a
constraint based upon the difference in the sum of vapor and liquid mole fractions, [EQ
8.83]. Rather than iterating to find the vapor fraction, V , or liquid fraction L = 1 – V ,
from [Ref. 1], V is set to 0 (zero - bubble point) or 1 (one - dew point) and we vary
pressure until the set of conditions is satisfied.
Equation of state formulation

The fugacities and Z -factors used in the flash calculations are obtained from the
equation of state.
The four equations of state implemented in PVTi are:
• (RK) Redlich-Kwong
• (SRK) Soave-Redlich-Kwong
• (ZJ) Zudkevitch-Joffe
• (PR) Peng-Robinson

Equation of state
The equations of state listed above are implemented using the formulation presented
by Martin, [Ref. 2], and Coats, [Ref. 3].
The equation of state for a real fluid is:
PV = nRTZ [EQ 8.88]
where P is the pressure, V the volume, n the number of moles, R the universal gas
constant, T the temperature and Z is obtained from the solution of the cubic equation:
3 2
Z + E2 Z + E1 Z + E0 = 0 [EQ 8.89]
with
E 2 = ( m 1 + m 2 – 1 )B – 1 [EQ 8.90]
2
E 1 = A – ( 2 ( m 1 + m 2 ) – 1 )B – ( m 1 + m 2 )B [EQ 8.91]
and
2
E 0 = – [ AB + m 1 m 2 B ( B + 1 ) ] [EQ 8.92]
The coefficients m 1 and m 2 depend upon the equation used:
Table 8.8 Equation of State coefficients
Mnemonic Equation of state m 1 value m 2 value
RK Redlich-Kwong 0 1
SRK Soave-Redlich-Kwong 0 1
ZJ Zudkevitch-Joffe 0 1
PR Peng-Robinson
1+ 2 1– 2
The cubic equation for the Z -factors may be solved to obtain Z -factors for liquid and
vapor phases. Generally three solutions are obtained. The distinction between the
liquid and vapor phase is then made by choosing the smallest root as the Z -factor for
the liquid phase and the largest root as the Z -factor for the vapor phase.
Fugacity coefficients are calculated using:
⎛ fi ⎞ A 2S B ( Z + m2 B ) Bi
ln ⎜ -------⎟ = – ln ( Z – B ) + ----------------------------- --------i – -----i ln ------------------------- + ----- ( Z – 1 ) [EQ 8.93]
⎝ px i⎠ ( m 1 – m 2 )B A B ( Z + m1 B ) B
where
Si = ∑ Aij x j [EQ 8.94]
n n
A = ∑ ∑ ( x j x k A jk ) [EQ 8.95]
j = 1k = 1

Equation of state
n
B = ∑ ( xj Bj ) [EQ 8.96]
j=1
1---
2
A jk = ( 1 – k jk ) ( A j A k ) [EQ 8.97]
and k jk are binary interaction coefficients, normally between hydrocarbons and non-
hydrocarbons.
These four equations, [EQ 8.94] to [EQ 8.97], express the mixing laws used in all the
equations of state.
The variables A j and B j are defined by the following equations:
p rj
A j = Ω A ( T, j ) ------- [EQ 8.98]
2
T rj
and
p rj
B j = Ω B ( T, j ) ------- [EQ 8.99]
T rj
Ω A ( T, j ) and Ω B ( T, j ) used in equations [EQ 8.98] and [EQ 8.99] are functions of the
acentric factor ω and the reduced temperature T rj .
For Redlich-Kwong
– 1---
2
Ω A ( T, j ) = Ω A T rj [EQ 8.100]
0
Ω B ( T, j ) = Ω B [EQ 8.101]
0
For Soave-Redlich-Kwong
1--- 2
⎛ ⎞
Ω A ( T, j ) = Ω A 1 + ( 0.48 + 1.57ω j – 0.17ω j ) 1 – T rj⎟
2 ⎜ 2
[EQ 8.102]
0 ⎜ ⎟
⎝ ⎠
Ω B ( T, j ) = Ω B . [EQ 8.103]
0
For Zudkevitch-Joffe
– 1---
2
Ω A ( T, j ) = Ω A F aj ( T )T rj [EQ 8.104]
0
Ω B ( T, j ) = Ω B F bj ( T ) [EQ 8.105]
0

Equation of state
For Peng-Robinson
1--- 2
⎛ ⎞
Ω A ( T, j ) = Ω A 1 + ( 0.37464 + 1.54226ω j – 0.2669ω j ) 1 – T rj⎟
2 ⎜ 2
[EQ 8.106]
0 ⎜ ⎟
⎝ ⎠
Ω B ( T, j ) = Ω B [EQ 8.107]
0
The normal PR form can be optionally modified for large acentric factor, using the
2 3
factor ( 0.379642 + 1.48503ω j – 0.164423ω j + 0.016666ω j ) rather than
2
( 0.37464 + 1.54226ω j – 0.2699ω j ) for ω j > 0.49 . This correction is invoked by use of the
PRCORR keyword or interactively from the Equation of State panel. For further
information see "The fluid model" on page 109.
Ω A and Ω B are constants depending upon the equation of state:
0 0
Table 8.9 Equation of State constants
Equation ΩA ΩB
0 0
RK, SRK, ZJ 0.4274802 0.086640350

PR 0.457235529 0.077796074
Note In the program, these default values may be over-written on a component

basis either interactively, or by the use of the OMEGAA and OMEGAB keywords.
Zudkevitch-Joffe equation
The Zudkevitch-Joffe equation contains additional temperature factors, denoted
F aj ( T ) and F bj ( T ) , multiplying the usual RK temperature dependence. These are
adjusted to match the pure-component fugacity values along the vapor pressure line,
and to observe the correct component liquid density.
In the program the required variation of component saturation pressure and liquid
density with temperature are obtained using correlations of Reidel, and Gunn and
Yamada. These correlations require the input of the normal boiling point and the
reference liquid density at a specified reference temperature.
Surface tensions
The surface tension between the liquid and vapor phase of a multi-component mixture
can be estimated by the Macleod-Sugden relationship, [Ref. 18]:
1--- Nc
4 Liq Vap
σ mix = ∑ [ Pi ] ( ρm xi – ρm yi ) [EQ 8.108]
i=1

Equation of state
th
where [ P i ] is the parachor of the i component, see [Ref. 18], which has a liquid and
vapor mole fraction of x i and y i respectively, and the liquid and vapor molar densities
Liq Vap
are ρ m and ρ m respectively. If the parachors are assigned values consistent with
3
[Ref. 18] and the molar volumes are expressed in gmoles/ cm then the surface tension
σ have units of dynes/cm.
Three-parameter equation of state

The traditional weakness of the so-called two-parameter equation of state, such as the
Peng-Robinson, Redlich-Kwong, etc., above, is their poor prediction of liquid
properties, especially liquid densities and saturations.
Peneloux et al., [Ref. 26], proposed a molar volume correction for the SRK Equation of
State, which is also applicable to any cubic Equation of State. This technique, referred
to as volume translation, adds a third parameter to the Equation of State, which greatly
improves liquid properties estimations.
Note In PVTi, this correction is available for the Peng-Robinson and the Soave-
Redlich-Kwong Equation of State.
For a mixture of N components, the phase volume, V mol,p , is given by:
N
EoS
V mol,p = V mol,p – ∑ zi ci [EQ 8.109]
i=1
where:
EoS
• V mol,p is the molar volume of the phase
• p = (liquid, vapor) predicted by the traditional (two-parameter) Equation of

State
• z i = ( x i ,y i ) are the liquid and vapor mole compositions
• c ri constitute a set of volume corrections.
The component corrections are usually related to the set of dimensionless shift
parameters, s i , by:
c ri
s i = ------ [EQ 8.110]
bi
where:
RT ci
b i = Ω b, i ----------- . [EQ 8.111]
p ci
See [Ref. 26].

Equation of state
Shift parameters
No Temperature Dependence (“None”)

In the simplest case, where the Volume Shift dependence is “None”, the Shift
parameters for the inorganic and light hydrocarbon components are calculated using
the temperature-dependent correlations given by Søreide, [Ref. 34].
The functional form of these polynomials depends on whether the PR or SRK EOS is
being used and on whether the reduced temperature, Tr (= T ⁄ T crit ) is less than or
greater than 1. For this simple case, where there is no temperature dependence on the
Volume Shifts a value of Tr=0.7 is used in the equations.
In the case of the PR EOS the equation take the form:
1.355
s i = T r – 0.741 – 0.164 + 0.479ω i + 0.428 exp [ 25.33 ( T r – 1 ) ] [EQ 8.112]
and for the SRK EOS:

3.713
s i = T r – 0.904 + 0.587ω i + 0.369 exp [ 10.369 ( T r – 1 ) ] [EQ 8.113]
where, for each component, s i is the dimensionless Volume Shift, ω i the acentric
factor and, as mentioned above, T r =0.7.
For heavier components ( M w > 90 ) the shifts are estimated on a component by

component basis as the difference in the calculated single component molar volume
using the two-parameter Equation of State at the components reference pressure and
temperature (often 60 °F and 14.7 psia), and the reference molar volume (mole weight
divided by reference density).
The shift parameters of the heavier components are potential regression variables and
are added to the allowed set if you request either the PR3 or the SRK3 Equation of State.
Experience has shown that simultaneous matches to saturation pressure and density
can be achieved more easily as a result, and therefore the use of these Equations of State
in PVTi is recommended.
One obvious deficiency of this model is that the shift coefficients, c i , are only valid at
one temperature. If one attempts to model the change of molar volume of a pure
component, such as Hexane, with temperature using, say, the PR3 Equation of State,
and compares that against the known thermal expansion coefficient, see [Ref. 18], one
will find a discrepancy.
In an attempt to overcome this deficiency, two options have been added to PVTi
whereby temperature-dependence is introduced using differing methods which are
now outlined.
Linear Expansion Only

In this model, the volume shifts of all the components are modified with a linear
thermal expansion coefficient, κ , using the equation:
c i ( T ) = c ri ( T ri, P ri ) [ 1 + κ i κ ( T – T ri ) ] [EQ 8.114]

Equation of state
where c ri ( T ri, P ri ) for a particular component is the dimensional volume shift
specified in equation [EQ 8.109]. Physically these c ri coefficients correspond to the
difference in volume of 1 mole of a particular component at temperature T and
pressure P (pressure at standard conditions - usually 60F), and 1 mole of that
component at its reference temperature and pressure, T ri and P ri .
The κ i coefficients are determined using a formula which is a fitted functional (of mole
weights) of the first ten paraffins from [Ref. 17] and is given by:
κ i = 0.9 – 0.002MW i, MW i < 200 [EQ 8.115]
κ i = 200, MW i ≥ 200
–1
κ is a special regression function in PVTi, and takes the default value of 0.0005 °C ,
determined from a fit to a crude oil. Physically, it represents a thermal expansion factor,
and has keyword THERMX within PVTi.
Polynomial Correlations
If the temperature dependence of the Volume Shifts is set to Polynomial Correlations
then the equations of Soreide are used which were introduced in the case of Tr<1 in
equations [EQ 8.109] and [EQ 8.109]. When the None option was set we always set
Tr=0.7 and so these equations were fine. However, if a temperature above Tc is supplied
to the algorithm with the Polynomial Correlations option set so that Tr>1, then PVTi
chooses from a set of equations depending on the mole weight of the component. These
equations can be found in [Ref. 34].
Note The Soreide correlations only cover light components up to C 6 .
For heavier components ( M w > 90 ) the thermal expansion scheme outlined in the
"Linear Expansion Only" on page 336 is applied. However, since the scheme is now
applied to regions where the fitted functional κ i is no longer valid, κ i is set to unity,
and the thermal expansion factor, κ , applies directly. For this reason it is often a good
choice as a regression variable for heavy hydrocarbon components if one is finding
difficulty matching liquid densities.
The Linear Expansion Only model has some advantages in that since the modification
to volume shifts is linear, the volume shifts may be regressed for all components.
For the Polynomial Correlations model, since the correlations of Søreide are not linear,
you cannot regress the shift parameters of lighter components to which the correlations
have been applied.
Hint However, an inspection of the behavior of volume shifts with temperature, as

given in [Ref. 36], shows that volume shifts generally do not obey a linear
relation with temperature, and for this reason, the second method often gives
better results and is the recommended choice.

Equation of state
Note If the volume shifts are Dependent, then they are calculated so that the
calculated reference density (the density at the reference temperature and
standard conditions pressure) is guaranteed to match the observed value. This
means that the volume shifts are dependent on the critical properties and
acentric factor of the component. If the volume shifts are Independent this
guarantee is broken and the volume shifts can be varied during regression. In
this situation they are not affected by changes in the critical properties or
acentric factors.
Multiphase flash
Standard algorithm
The conventional two phase flash calculation starts with a stability test for the feed
composition, based on Michelsen’s (1982) stability criterion, . If unstable, a flash phase
split calculation is performed to determine the two phase compositions. The
multiphase flash continues in a step-wise fashion: if any one of the current N phases
exhibits instability, an N+1 phase flash split calculation is performed, the resulting
tested for stability. The maximum allowed number of phases is five, after that is reach
no more phases are split off.
Options with the Multiphase flash

The Multiphase flash has three modes:
1 Conventional two phase mode.
2 Three phase mode, one of which is restricted to water.
3 The full multiphase flash mode.
This is the default. In this mode there are now two different ways of performing a
multiphase flash calculation.
• The standard method.
This method is the one that was in use for the 2002A release and initially
assumes a single phase equal to the feed. Stability checks are performed and
subsequent phases are split off if the system is unstable.
• Instead of the flash starting with the feed an initial guess/guesses can be used
assuming there are 4 phases present, which are initially undetermined.

Equation of state
Stability analysis
The stability of each phase present in a multiphase regime can be tested with the
Michelsen analysis, [Ref. 16], precisely as for the two phase case, using the Gibbs
energy tangent plane distance criterion. However, the initial estimates for phases that
will be split off are not limited to the vapor/liquid equilibrium phases characterised by
the Wilson K-values: additional estimates are chosen from the following phases:
• Each inorganic component
• Lightest hydrocarbon
• Heaviest hydrocarbon
• Ideal gas
• Arithmetic mean of existing phases.
• Wax - prediction of cloud point - SPE 27629
The solutions method used for Michelsen stability criterion is a Successive substitution
with GDEM (general dominant eigenvalue method) follow by BFGS if that does not
work. These trial phases are tested in turn until one indicates instability, in which case
the whole system is unstable, and the algorithm proceeds to split of that phase,
provided that phase does not already exist. If none of the trial phase indicates
instability, the state is assumed to be stable.
Phase Split Calculations

The multiphase flash calculations are based on Gibbs energy minimisation of the
system, and is fully defined by the multiphase flash equation and the Gibbs energy
stationary point criterion.
Systems of Equations
Material Balance, Mass conservation equations or consistency equations

Consider a hydrocarbon system consisting of one mole of mixture or feed of
composition { z i }. In general, this will consist of { β m } moles from each of the M phases,
with composition { z im }, at some pressure P and temperature T.
Mass conservation gives:
∑ βm = 1 [EQ 8.116]
∑ zim = 1 [EQ 8.117]
∑ βm zim = zi [EQ 8.118]
m
These consistency rules can be checked at the end of the multiphase flash calculation.
Defining the multiphase equilibrium ‘K’-values as:

Equation of state
z ij
K ij = -----------
z ir ( i )
Where r ( i ) is the reference phase index chosen for component i, usually so that the
largest amount of component i is in the reference phase.
Gibbs Energy Stationary Conditions

The solution of the multiphase flash satisfies the consistency equations and is the
absolute minimum of the Gibbs free energy.
The reduced Gibbs energy of an M phase system is written as:
G-
g° = ------
RT
= ∑ ∑ βm zim ln ( fim ) [EQ 8.119]
i m
th th
where f im is the fugacity coefficient in the i component and the m phase. Where
the fugacties are calculated from the equation of state.
The first derivative with respect to the scaled mole numbers Θ jk = β k z jk ⁄ z j can be
written in various forms:
g· ° = z j ( ln f jk – ln f jr ( j ) ) [EQ 8.120]
= z j ( ln z jk – ln z jr ( j ) + ln Φ jk – ln Φ jr ( j ) ) [EQ 8.121]
= z j ( ln K jk + ln Φ jk – ln Φ jr ( j ) ) [EQ 8.122]
where Φ jk is the fugacity coefficient of component j in the k’th phase. In equilibrium,

the Gibbs energy is at a minimum, and the first order stationary condition thus implies
fugacity balance between like components in all phases.
The second derivative of the Gibbs free energy is:
⎛ δ mk – δ mK⎞ ⎛ δ ij ∂ ln Φ im ∂ ln Φ im⎞
g·· ° = z i z j ⎜ -------------------------⎟ ⎜ ------- - – ∑ z pm -------------------⎟
– 1 + ------------------
⎝ βm ⎠ ⎝ z im ∂z im ∂z pm ⎠
p
⎛ δ Mk – δ MK⎞ ⎛ δ ij ∂ ln Φ iM ∂ ln Φ iM⎞
– ⎜ --------------------------⎟ ⎜ ------- - – ∑ z pM -------------------⎟
- – 1 + ------------------
⎝ βM ⎠ ⎝ z iM ∂z jM ∂z pM ⎠
p
where K = r ( k ) , M = r ( m ) and δ ij is the Kronecker delta and the fugacity

coefficient derivatives are calculated from the Equation of State being used.

Equation of state
Finding Stationary solutions
Starting from initial conditions provided by the stability analysis (or input directly),
the general algorithm is repeating cycle of the following:
• iterate forward to new J = ln ( K ) values

• solve the multiphase flash equation for new beta-values
• update the compositions. Checking to see solution is consistent, that is mole
fractions have to add up to one.
The algorithm has two attempts at converging to a local minimum. First method
updates with direct or successive substitution with GDEM of the J-values. If it does not
converge that is the fugacity balance equation residuals fails to meet a specified
tolerance, it proceeding to a try continuing the convergence with Newton-
Raphson/Murray method.
Successive Substitution (SS)

The SS part incorporates the General Dominant Eigenvalue Method (GDEM)
acceleration algorithm. For details see Crowe and Nishio. The variable used are the J-
values themselves, updated from the second form of the first derivative given in
equations [EQ 8.120], [EQ 8.121] and [EQ 8.122]. The reference phase for each
component is chosen as the phase with the highest mole fraction of that component, so
as to scale each K value close to and below one.
Newton-Raphson (NR)
The NR part uses a Murray decomposition of the Gibbs Hessian to ensure positive
definiteness of the Jacobian, and therefore progress towards a minimum. Occasionally,
a cubic trace back augmented line search in the Newton direction indicated is required
when the Gibbs free energy is found not to decrease.
The variables used are the scaled mole numbers, and the reference phase is the same as
in the SS, that is, the phase with the highest mole fraction.
Trace Elements
If a phase possesses a near-zero component composition, then the variable for that
component and phase can be dropped from the variable list (that is not changed in the
SS or NR) and its value calculated after each iteration by invoking fugacity balance
between it and the fugacity of a variable that was update in the usual way. In doing so,
the trace element fugacity coefficient is calculate by setting its composition to zero. If
the component transcends the numerical zero boundary, its variable is reintroduced
into the variable set ready for the next iteration.
4-Phase Guess Method

Instead of doing a phase split method we can used initial guesses to obtain a different
initial starting points for the find stationary solutions using SS and NR, which might
converge to a different stationary state. See Trebble ([Ref. 63]).

Equation of state
There are various initial guesses that can be implemented.
• Wilson initiation, proposed by Wilson ([Ref. 1])
• Wilson and Antonie
• NISA (Non-Iterative Stability Analysis)
• Wilson and Stability
• NISA and Stability
• Liquid-Liquid Initiation
• Trebble’s Vapor-Liquid-Liquid Initiation
• Liquid-Vapor and Wax
• Wax and Liquid-Liquid-Vapor Initiation.
• Five phase initial guess.
It was found that for a flash with wax that the wax initial guess did give a lower Gibbs
free energy some of the time. For the wax initial guess we assume the composition of
the liquid phase is the same as the feed. We then workout based on the wax, heavier
liquid and vapor K values by:
o
ln y i = ln x i + ln Φ i, V ( x i ) – ln Φ i, L [EQ 8.123]
o
ln x 2i = ln x i + ln Φ i, L ( x i ) – ln Φ i, L [EQ 8.124]
o
ln x 3i = ln x i + ln Φ i, W ( x i ) – ln Φ i, L [EQ 8.125]
o
where Φ denotes pure fugacity, L,V and W denote liquid, vapor and wax. y i
component in the vapor phase. x i , x 2i and x 3i are the components in the lightest
liquid, the heavier liquid and the wax phase.
Both result using the initial guess and standard algorithm can pass the stability test, so
in the multiphase where it is possible for wax to form, we do both the four phase wax
initial guess and the standard multiphase flash and compare results, taking the one
with the lowest Gibbs free energy.
In the case of the debug option switching on all the initial guesses and in the case for
the wax possible case fail to give an answer which pass the stability test, then all initial
guesses are tried and compared and the result with the minimum Gibbs free energy
taken as the answer.
Treatment of water systems

Water systems need to be given special treatment because of the inapplicability of the
Equations of State generally used for any phase, liquid or vapor, containing water. The
current lines of attack at this problem are
• Lattice-gas models
• local composition models
• different mixing rules
• use of henry’s law rather than an Equation of State
• modification of commonly used Equations of State

Equation of state
Lattice-gas models and local composition models have attracted attention through
their strong theoretical basis, incorporating hydrogen-bonding and electrolyte
thermodynamics respectively. Relaxation of the commonly used symmetry condition
for τ ij s yielding different mixing rules, is attractive for similar reasons. As yet though,
these approaches have found little application, and are not easily incorporated into
existing Equation of State packages.
Henry’s Law, based essential on ‘look-up’ fugacites for given solutes and solvent, has
had a longer history of application to engineering ‘rule-of-thumb’ problem, but is
difficult to incorporate in a GEM multiphase algorithm, requires significant amounts
of data, uses non-exact models for pressure and salinity corrections, and is restricted to
specific solvents and solutes.
The Valderrrama-Patel-Teja Equation of State is a 3 degree Equation of State that has
extra terms involving the compressibility factor, and has had some success in dealing
with water systems. Again, this has not had wide use, and again, has still to be
corrected for salinity. The most common approach in the literature is to use the more
popular Equations of State, but with some modification, particularly to the τ ij s used.
The Peng-Robinson Equation of State has received this treatment, using T-dependent
τ ij s between water and most SCN groups. The ‘PRW’ Equation of State is used here
details are given in "Peng-Robinson and Water Equation of State" on page 344.
The multiphase flash allow three separate treatments for systems containing water that
produce a water phase, each based on the PRW Equation of State:
• maintain the water phase as pure
• allow CO 2 only to dissolve

• allow all components to dissolve
If the given system contains water the stability analysis attempts specifically to split off
a water phase before any other. A full Michelsen stability analysis is not required:- each
of the above options describe a water-dominated phase, so it is enough to compare the
activity of water in a pure water phase with that in the feed mixture, and reintroduce
the impurities required by the option at a later stage. In terms of fugacities, a pure
water phase will split off if and only if fugacity of water in the feed is greater than
fugacity of a pure water ( f w ( feed ) > f w ( pure ) ) or:
ln ( Φ w ) + ln ( z w ) > ln ( Φ w )
feed feed pure
where the w denotes the component index of the water. The difference between the
options is then implemented in the ‘bring-back’ treatment for trace components used
in the phase split calculations, either none, CO 2 only, or all.

Equation of state
Water- CO 2 bic-fitting
A particular modification to the use of PWR Equation of State is made when only CO2
is dissolved in water, in allowing for salinity effect CO 2 solubility decrease with
salinity, but the PRW Equation of State makes no allowance for this. It has been
possible, however, to establish a polynomial form for the H 2 O – CO 2 τ ij by iterating
adjusting (with a NR scheme) the H 2 O – CO 2 τ ij until the solubility predicted by
the flash calculation matches experimentally determined solubilities. With a set of such
τ ij s one can fit a 3 degree surface polynomial in p, T, S (salinity):
τ H O – CO = f 1 P, T ) + Sf 2 P, T )
2 2
2
f i = a 1i + a 2i T + a 3i P + a 4i TP + a 5i T
2 2 3 3
+a 6i P + a 7i T P + a 9i T + a 10i P
where a linear correction for salinity is assumed.
Peng-Robinson and Water Equation of State

The treatment of water systems is based on the PRW (Peng-Robinson and Water)
equation of state. The original PR Equation of State has the form (for the usual
symbols):
a(T)
RT - – ------------------------------------------------
P = ----------- -
V – b V(V + b) + b(V – b)
1⁄2
a = ∑ ∑ xi xj ( 1 – δij ) ( ai a j )
i j
b = ∑ xi bi
i
a i = a ci α i
2 2
R T ci
a ci = 0.45724 ----------------
P ci
1⁄2 1⁄2
α = 1 + κ ( 1 – Tr )
2
κ = 0.37464 + 1.54226ω i – 0.26992ω i
RT ci
b i = 0.07780 -----------
P ci
Two modification are made for water systems. Firstly, for the water component in any
phase, a correction is made to improve predicted water vapor pressures:

Equation of state
1⁄2 1⁄2
α = 1.0085677 + 0.82514 ( 1 – T r )
1⁄2
when Tr < 0.85
Secondly, in the aqueous liquid phase, a temperature-dependent interaction parameter
is used, so that the mixing rules change slightly to:
1⁄2
a = ∑ ∑ xi xj ( 1 – τij ( T ) ) ( a i a j )
i j
when T r ≥ 0.85 , or the phase is not the aqueous liquid phase, the original PR
Equation of State is used.
The temperature τ ij s required may be found from the published curves given by Peng
& Robinson. The following forms have been derived:
τ C ↔ H O = 1.659T r P rw – 0.761
1 2
τ C ↔ H O = 2.109T r P rw – 0.607
2 2
2 2
τ C ↔ H O = – 18.032T r P rw + 0.9441T r P rw – 1.208
3 2
τ C ↔ H O = 2.800T r P rw – 0.488
4 2
⎛ T c⎞
τ C ↔ H O = ⎜ 0.4 ------⎟ T r P rw – 0.8 , n > 4
n 2 ⎝ P c⎠
τ N ↔ H O = 0.402T r – 1.586
2 2
τ H S ↔ H O = 0.22T r – 0.19
2 2
2
τ CO ↔ H O = – 0.074T r + 0.478T r – 0.503
2 2
The general C n form, was fitted here.
Viscosity correlations
Three correlations are available in PVTi for the estimation of viscosities, namely that
due to "Lohrenz, Bray and Clark" on page 345, [Ref. 7], the method due to "Pedersen et
al." on page 347 (PED), [Ref. 5] and [Ref. 36], and also that of "Aasberg-Petersen et al"
on page 349, [Ref. 64].
Lohrenz, Bray and Clark

The most widely used correlation for the prediction of liquid and vapor viscosities in
reservoir simulators is that due to LBC.

Equation of state
The viscosity being related to a fourth-degree polynomial in reduced density,
ρr = ρ ⁄ ρc :
–4 1 ⁄ 4 2 3 4
[ ( η – η∗ )ξ + 10 ] = a1 + a2 ρr + a3 ρr + a4 ρr + a5 ρr [EQ 8.126]
where
a 1 = 0.1023000
a 2 = 0.0233640
a 3 = 0.0585330
a 4 = – 0.0407580
a 5 = 0.0093324 [EQ 8.127]
and η∗ is the low-pressure gas mixture viscosity. ξ is the viscosity-reducing

parameter, which for a fluid mixture is given by:
N 1⁄6 N –1 ⁄ 2 N –2 ⁄ 3
ξ = ∑ z i T ci ∑ z i M wi ∑ z i p ci [EQ 8.128]
i=1 i=1 i=1
The critical density ρ c is evaluated from:
⎛ N ⎞ –1
ρc = Vc = ⎜ ⎟
–1
⎜ ∑ ( z i V ci ) + z C
7+
V cC
7+ ⎟
[EQ 8.129]
⎝ i = 1, i ≠ C ⎠
7+
where the critical volume of the plus fraction is found from:

V cC = 21.573 + 0.015122M wC – 27.656γ C + 0.070615M wC γ [EQ 8.130]
7+ 7+ 7+ 7+ C 7+
The dilute gas mixture viscosity is as given by Herning and Zippener, [Ref. 28]:
N N –1
1⁄2 1⁄2
η∗ = ∑ z i η i∗ M wi ∑ z i M wi [EQ 8.131]
i=1 i=1
where the dilute gas viscosities of the individual components, η i∗ are derived from
expressions due to Stiel and Thodos, [Ref. 29]:
– 5 1 0.94
η i∗ = 34 ×10 ---- T ri T ri < 1.5 [EQ 8.132]
ξi
–5 1 0.625
η i∗ = 17.78 ×10 ---- ( 4.58T ri – 1.67 ) T ri > 1.5 [EQ 8.133]
ξi
where
1⁄6 –1 ⁄ 2 –2 ⁄ 3
ξ i = T ci M wi p ci [EQ 8.134]

Equation of state
Hint By making the viscosity a function of the fourth power of density, results are
very sensitive to small differences in estimation of density. It is not unusual for
this technique to predict a viscosity only 50% of the measured viscosity. Small
changes in critical volumes or critical Z -factors remedy the error but it is
recommended that they are changed as a single group, from [EQ 8.130].
Pedersen et al.
Viscosities can be calculated from a modified form of the corresponding states method.
A group of substances obey the corresponding states principle if the functional
dependence of the reduced viscosity, η r , say, on reduced density and temperature, ρ r
and T r , say, is the same for all components within the group, namely:
η r ( ρ, T ) = f ( η r ,T r ) [EQ 8.135]
in which case comprehensive viscosity data is only needed for one component of the
group, which is denoted as the reference substance, to be given the subscript ( o ) all
other components are identified with the subscript ( x ).
Generally, the critical viscosity, η c , is not known but it can be estimated from the
inverse of [EQ 8.135]),
–1 ⁄ 2 2 ⁄ 3 –1 ⁄ 2
η r ( ρ, T ) = η ( ρ, T )T c Vc Mw [EQ 8.136]
Thus, the viscosity of component x at temperature T and a pressure with density ρ , is

given by:
1 ⁄ 2 –2 ⁄ 3 1⁄2
T cx V cx M wx
η r ( ρ ,T ) = ------------ --------------- -------------- η o ( ρ o ,T o ) [EQ 8.137]
T co V co M mo
where η o = ρρ co ⁄ ρ cx , T o = TT co ⁄ T cx and η o is the viscosity of the reference

substance at T o and ρ o .
Oil mixtures contain a range of components with molecular weights ranging from 16
(Methane) to about 1100 ( C 80 ). It cannot be expected for C 1 and C 80 to belong to a
group where a simple corresponding states principle applies. Generally, a third
parameter must be added to account for the shape of the molecules, such as the acentric
factor.
Pedersen et al., [Ref. 5], presented a corresponding states principle depending on
reduced temperature and reduced pressure, η r = f ( p r, T r ) where:
η
η r = ηξ = ------------------------------------------ [EQ 8.138]
–1 ⁄ 6 2 ⁄ 3 1 ⁄ 2
Tc pc Mw
The deviation from the simple corresponding states principle is expressed in terms of
a rotational coupling coefficient, denoted α , to give:

Equation of state
–1 ⁄ 6 2⁄3 1⁄2
T c, mix p c, mix M w, mix α mix
η mix ( p, T ) = ----------------------------- ------------------------- ----------------------------- ------------ η o ( p o, T o ) [EQ 8.139]
T co p co M wo αo
where
p co α o
p o = p ----------------------------- [EQ 8.140]
p c, mix α mix
T co α o
T o = T ----------------------------- [EQ 8.141]
T c, mix α mix
The critical temperature and volume for unlike pairs of molecules are given by:
1⁄2
T c, ij = ( T c, i T c, j ) [EQ 8.142]
1 1⁄3 1⁄3 3
V c, ij = --- ( V c, i + V c, j ) [EQ 8.143]
8
where the critical volume of a component can be expressed in terms of the critical
temperature and pressure using the real gas law and the critical Z -factor.
Assuming a constant Z c for all components,[EQ 8.143] becomes:
1--- ⎛ ⎛ T c, i⎞ 1 ⁄ 3 ⎛ T c, j⎞ 1 ⁄ 3⎞ 3
V c, ij = constant ⎜ ⎜ ----------⎟ + ⎜ ----------⎟ ⎟ [EQ 8.144]
8 ⎝ ⎝ p c, i⎠ ⎝ p c, j ⎠ ⎠
The mixture critical temperature is found from:

N N
∑ ∑ z i z j T c, ij V c, ij
= 1j = 1
T c, mix = i-------------------------------------------------------
- [EQ 8.145]
N N
∑ ∑ z i z j V c, ij
i = 1j = 1
Combining [EQ 8.143] and [EQ 8.145] gives:
N N
⎛ T c, i⎞ 1 ⁄ 3 ⎛ T c, j⎞ 1 ⁄ 3 3 1⁄2
∑ ∑ i j ⎜⎝ --------- z z
p c, i ⎠
-⎟ ⎜ ----------⎟
⎝ c, j ⎠
p
[ T c, i T c, j ]
= 1j = 1
T c, mix = i----------------------------------------------------------------------------------------------------------------------------- [EQ 8.146]
N N
⎛ T c, i⎞ 1 ⁄ 3 ⎛ T c, j⎞ 1 ⁄ 3 3
∑ ∑ i j ⎜⎝ ---------z z
p c, i ⎠
-⎟ ⎜ ----------⎟
⎝ p c, j ⎠
i = 1j = 1
and the mixture critical pressure is evaluated from:
N N
⎛ T c, i⎞ 1 ⁄ 3 ⎛ T c, j⎞ 1 ⁄ 3 3 1⁄2
∑ ∑ i j ⎜⎝ ---------
8 z z
p c, i⎠
-⎟ ⎜ ----------⎟
⎝ c, j ⎠
p
[ T c, i T c, j ]
i = 1j = 1
p c, mix = --------------------------------------------------------------------------------------------------------------------------------
- [EQ 8.147]
⎛ N N 1⁄3 T 1 ⁄ 3 3⎞ 2
⎜ ⎛ c, i⎞
T ⎛ c, j⎞ ⎟
⎜ ∑ ∑ i j ⎜⎝ p ⎟⎠ ⎜ ----------⎟
z z ----------
⎝ p c, j ⎠ ⎟
⎝i = 1j = 1 c , i ⎠

Equation of state
The mixture mole weight is given by:
–4 ⎛ 2.303 2.303⎞
M w, mix = 1.304 ×10 M – Mn + Mn [EQ 8.148]
⎝ w ⎠
where M w and M n are the weight average and number average mole weights,
respectively.
This mixing rule is derived empirically on the basis of available viscosity data and
assigns a larger influence to heavier components.
The α -parameter for the mixture is found from:
– 3 1.847 0.5173
α mix = 1.0 + 7.378 ×10 ρ r M w, mix [EQ 8.149]
where the α of Methane, the reference substance, is given by:

1.847
α 0 = 1.0 + 0.031ρ r [EQ 8.150]
Note The constants and exponents in equations [EQ 8.149] and [EQ 8.150] have been
found from fitting to experimentally determined viscosity data.
It has been our experience that the method of Pedersen gives much better prediction of
viscosities than does the method of Lohrenz, Bray and Clark. Neither method is
particularly good but typical errors based on un-regressed Equation of State data are,
LBC ∼ 50 % and PED ∼ 90 % of the reported values. However, you should treat
experimentally measured viscosities with some caution, as viscosities are often
calculated.
Aasberg-Petersen et al
The Aasberg-Petersen model is also based on the principle of corresponding states, but
uses two references fluids, methane and decane, instead of just the one (methane) in
the case of the Pedersen et al. model. The usual problem with the Pedersen model is
that inaccurate predictions occur for fluids with components that are significantly
different in molecular weight to the reference component (methane).
The idea of this model is to use a heavier second reference component and effectively
create an optimum reference component by using the molecular weight of the fluid as
an interpolation parameter between the two reference components. Decane is chosen
as the second reference component because it is the heaviest alkane for which a
significant amount of experimental viscosity data is known. Methane is a natural
choice as the first reference component because of its presence in large mole fractions
in most reservoir fluid mixtures
The interpolation law is used to calculate the reduced viscosity of the optimum
reference component (denoted with the subscript x) using the reduced viscosities of
methane and decane and is obtained using the following expression:
MW x – MW 1 ⎛ η r2⎞
ln η rx = ln η r1 + -------------------------------- ln ⎜ --------⎟ [EQ 8.151]
MW 2 – MW 1 ⎝ η r1⎠

Equation of state
MW is the molecular weight and subscripts 1 and 2 refer to the reference components.
The functional form of equation [EQ 8.151] was originally suggested by Teja and Rice
(1981), [Ref. 65], using the acentric factor instead of MW. This is not possible in the
present work, since the acentric factor decreases with increasing molecular weight for
heavy oil fractions.
The reduced properties are determined from:
E
E r = ------, E = T, P, η [EQ 8.152]
Ec
Subscripts r and c indicated reduced and critical properties respectively. The following
expression is used to evaluate the critical viscosity (Pedersen et al. 1989):
1 2 1
--- --- – ---
2 3 6
ηc = C ⋅ MW P c T c [EQ 8.153]
C is a constant. From equations [EQ 8.151] and [EQ 8.142][EQ 8.142] the following
equations may be derived for determination of the viscosity:
η cx η 1 ( T 1, P 1 ) η 2 ( T 2, P 2 )η c1 K
η x = ----------------------------------- ------------------------------------ [EQ 8.154]
η c1 η 1 ( T 1, P 1 )η c2
MW x – MW 1
K = -------------------------------
- [EQ 8.155]
MW 2 – MW 1
η 1 and η 2 are evaluated at conditions corresponding to the reduced temperature and
pressure of component x:
T ci T
T i = -----------, i = 1, 2 [EQ 8.156]
T cx
P ci P
P i = -----------, i = 1, 2 [EQ 8.157]
P cx
The model outlined in the above is extended to mixtures using the same mixing rule as
the Pedersen et al. model, that is equations [EQ 8.146] and [EQ 8.147]. The mixture
molecular weight is calculated using the formula:
1.56079 1.56079
MW mix = MW n + 0.00867358 ( MW w – MW n ) [EQ 8.158]
where [EQ 8.159]
N
2
∑ z i MW i
=1
MW w = i---------------------------
-
N
∑ z i MW i
i=1
and

Equation of state
N
MW n = ∑ z i MW i [EQ 8.160]
i=1
The constants in equation [EQ 8.142] are determined by regression using experimental
viscosity data for binary mixtures and oils.
The following equations are used for determination of the viscosity of the reference
components:
η 0 = η k ( T ) + ρη 1 ( T ) + η 2 ( ρ, T ) [EQ 8.161]
9 (i – 4)
---------------
3
ηk ( T ) = ∑ GV i T [EQ 8.162]
i=1
2
η 1 ( T ) = A + B C – ln ⎛ ---⎞
T
⎝ F⎠
[EQ 8.163]
j
η 2 ( T, ρ ) = H 2 exp ⎛ j 1 + ---4-⎞
⎝ T⎠
[EQ 8.164]
0.1 ⎛ j3 ⎞ 0.5 ⎛ j6 j7 ⎞
H 2 = – 1 + exp ρ ⎜ j 2 + ---------⎟ + θρ ⎜ j 5 + ---- + ------⎟ [EQ 8.165]
⎝ T ⎠
1.5 ⎝ T T 2⎠
where [EQ 8.166]
( ρ – ρc )
θ = -------------------
ρc
These equations were originally developed by Hanley et al. (1975), [Ref. 66], to
correlate the methane viscosity. For methane the values of the GVi parameters in
equation [EQ 8.152] given by Hanley et al. were maintained. For decane GV4-GV9 are
equal to zero and GV1-GV3 were taken from the DIPPR tables (1985). All the
parameters in equations [EQ 8.59]-[EQ 8.165] were estimated. In Table 8.10, data of the
parameter estimation are given and Table 8.11 lists all parameters for the pure
component viscosity correlation. Of the parameters in Table 8.11 are used the viscosity
is obtained in μP if the density is given in g/cm3.
Table 8.10 Parameter estimation data. N is the number of experimental points
Methane Decane
N 881 252
T-range (K) 91-523 244-477
P-range (atm.) 0-680 0-1000
Dev. (%) 3.1 3.8

Equation of state
Table 8.11 Parameter Values for Pure Component Viscosity Correlation
Methane Decane
GV1 -209097 0.2640
GV2 264276 0.9487
GV3 -147282 71.0
GV4 47164 0.0
GV5 -9491.9 0.0
GV6 1220.0 0.0
GV7 -96.28 0.0
GV8 4.274 0.0
GV9 -0.0814 0.0
A.100 23946 0.00248

B 343.79 81.35
C 0.4487 5.9583
F 168.0 490.0
j1 -22.768 -11.739
j2 30.574 16.092
j3 -14929 -18464
j4 1061.5 -811.3
j5 -1.4748 1.9745
j6 290.62 898.45
j7 30396 119620
The density of the reference components as a function of the pressure and temperature
is also required in equations [EQ 8.48] and [EQ 8.165]. For methane, as with the
Pedersen model, the 33-parameter MBWR-equation given by McCarthy (1974), [Ref.
67], was used to calculate the density. For decane we decided to used the flash available
within PVTi to obtain the density at a given pressure and temperature.

Equation of state
Finally, the following physical properties for methane and decane were used in Table
8.12 below when required.
Table 8.12 Physical Properties of Methane and Decane
Methane Decane
Tc 190.55 617.40
Pc 45.39 20.18
MW 16.043 142.284
ω 0.008 0.484
ρc 0.1649 0.2269
The predictions of the Aasberg-Petersen model have been shown to agree well with
experimental data over large pressure and temperature ranges. In particular this model
is able to handle mixtures with CO2, paraffinic and aromatic components with better
accuracy than the Pedersen model. A better match is also obtained than the Pedersen
model for heavier oils, where the size and shape of the molecules differ substantially
from the single reference component (methane).
Note It should be noted that the Aasberg-Petersen model is not recommended for
fluids with large concentrations of napthalenic components.

Strictly, binary interaction coefficients are interpreted as accounting for polar forces
between pairs of molecules.
Another interpretation is they account for tertiary and higher-order interactions in the
cubically (in volume) truncated form of the Second Virial Equation:
RT a
p = ------- + ------ + … [EQ 8.167]
V 2
V
Eitherway, they might be thought of as the fiddle-factors for the equation of state.
Strictly, they should be determined for all possible binary mixtures of hydrocarbons
and non-hydrocarbons by tuning their value to get a match between experimental and
theoretical (Equation of State) behavior. Thus, each new Equation of State requires that
a completely new set of binaries be developed: a laborious task indeed. This is the main
reason why the Soave-Redlich-Kwong and Peng-Robinson Equations of State tend to
dominate in reservoir applications.
Caution Many authors have suggested that binaries are the obvious Equation of
State parameter to adjust to match Equation of State to laboratory results,
especially the Methane to plus-fraction binary. However, Pedersen et al.,
[Ref. 30], have shown that this is problematic.
Given that our preferred Equation of State is the Peng-Robinson, we derive binaries
from one of two sources.

Equation of state
The current default is to assume they are zero for all hydrocarbon-hydrocarbon
interactions except for Methane to heavier hydrocarbons that vary like:
K C , j = 0.14γ j – 0.06 [EQ 8.168]
1
where γ j is the (liquid) specific gravity of the heavier component. In addition, the
hydrocarbon to non-hydrocarbon interactions take certain fixed values, see [Ref. 4].
The alternative set, activated by the appropriate option switch is a set due to Cheuh
and Prausnitz, see [Ref. 32].
The hydrocarbon to non-hydrocarbon interactions are as above but the hydrocarbon to
hydrocarbon binaries are given by:
⎛ 2 ( V V ) 1 ⁄ 6⎞ 6
c, i c, j
K i, j = A 1 – ⎜ --------------------------------------⎟ [EQ 8.169]
⎜ 1⁄3 1⁄3 ⎟
⎝ V c, i + V c, j ⎠
th
where V c, i is the molar volume of the i component and A is a special regression
variable which is generally in the range 0.15 ≤ A ≤ 0.25 .
There is some appeal in using the pre-multiplying coefficient to regress all binaries
together. This preserves symmetry, which might be lost using the Methane to plus
fraction binary approach of [EQ 8.169], which (see [EQ 8.169] and [Ref. 28]) can lead to
strange results.
These particular binaries are also given a temperature dependence by the use of an
additional multiplying factor that has the form:
1.0 + TC ( T – T std ) [EQ 8.170]
where TC has the fixed value of 0.0025.

Equation of state
Basic laboratory experiments
Introduction
This section describes the basic laboratory experiments and how PVTi attempts to
simulate them. More detail on experimental methods can be found in the excellent text
by Pedersen et al. [Ref. 30].
Schematic diagrams of the apparatus used for the standard laboratory techniques to be
described here can be found in the above text.
See "The fluid model" on page 109 and "Batch system and keywords" on page 163 for
further information on defining experiments in PVTi.
This section contains information on the following:
• "Blackoil systems" on page 355.
• "Gas condensate systems" on page 358.
• "Gas injection processes" on page 361.
• "Process simulation" on page 360.
Blackoil systems
Essentially two experiments are performed on black or crude oil systems:
• "Bubble point evaluation" on page 355
• "Differential liberation" on page 356 (sometimes referred to differential expansion).
Bubble point evaluation

Laboratory experiment
1 Having charged the PVT cell with the reservoir fluid, the system is left to come into
equilibrium at the required (reservoir) temperature and pressure or some higher
pressure where the fluid is a single phase liquid, whereupon the fluid volume is
noted.
2 The pressure is then dropped (at constant temperature) and the new fluid volume
is recorded. The bubble point pressure is then readily found from the discontinuity
in the pressure/volume relationship where the first bubble of gas is evolved from
the liquid. Since gases are more compressible than liquids, reduction in pressure
and liberation of gas results in an increase in rate of volume expansion of the (two
phase) fluid.
PVTi
1 PVTi finds the bubble point by taking one mole of the reservoir fluid and using
estimates for the K -values, constructs vapor compositions at a set of pressures,
testing for the stability of the second (vapor) phase using the method of Michelsen
[Ref. 16].

2 Once the highest-pressure, two-phase state has been found, the N equal fugacity
conditions and the one mole fraction constraint (see "Material balance" on page 330
and "Flash calculations" on page 330) are used to solve for the N + 1 unknowns of
N K -values and the bubble point pressure by iteration.
Having found the bubble point pressure, the crude oil would normally then be subject
to this experiment.
1 Starting from the bubble point pressure, the pressure is dropped by several
hundred psia, whereupon a volume of gas is evolved that is removed from the cell
at the (new) constant pressure. This process is repeated several times noting the
volume of gas evolved (at reservoir and surface conditions) and the volume of
liquid remaining.
2 The volume of liquid remaining at the last stage, which should be at standard or
atmospheric pressure, is then reduced to standard or atmospheric temperature and
re-measured. This liquid is discharged and its density is obtained. The density of
the liquid at the other pressure (and temperature) stages can be calculated from
mass balance knowing the volumes and molecular weight of the removed gas
streams.
The data usually quoted is:
• B o Oil formation volume factor
• B g Gas formation volume factor
• R s Gas-oil ratio
• ρ o Oil density
• γ g Gas gravity (density)
• Z g Gas deviation factor
• μ o Oil viscosity
• μ g Gas viscosity
where:
Vo ( p )
B o ( p ) = -------------
- [EQ 8.171]
std
Vo
Vg( p )
B g ( p ) = -----------------
- [EQ 8.172]
std
Vg ( p )

i = std
1 std
R s ( p ) = ---------- ∑ V g ( i ) [EQ 8.173]
std
Vo i = p
and
Mg( p )
γ g = ---------------
- [EQ 8.174]
M air
Here, V o , V g are volumes of oil and gas, M g , M air are mole weights of
hydrocarbon gas and air, the superscript std implies standard conditions, and the
summation for the gas-oil ratio is taken over the volumes of evolved gas from the
current pressure, p to the final pressure at std .
PVTi
1 In PVTi, the experiment is simulated by firstly locating the bubble point pressure,
as above. Then one mole of bubble point fluid is dropped in pressure and a flash
calculation is performed to determine the phase split and the volumes of oil and
gas. All the gas is removed and the liquid composition forms the feedstream for the
next pressure depletion stage, etc.
Note The "Definition of GOR in Diff. Lib." on page 159 program option allows three
further definitions of the GOR.
The first removes the stage to standard conditions and normalizes gas volumes to
the volume of oil at reservoir conditions, that is:
i = p last
1 std
R s ( p ) = -----------------------------
V o ( T, p bub ) ∑ Vg ( i ) [EQ 8.175]
i=p
where T is the temperature of the depletion experiment, and p last is the last
pressure stage specified.
2 The second definition specifies the GOR as an incremental one, that is:
std
Vg ( p )
R s ( p ) = -----------------
- [EQ 8.176]
std
Vo
3 The third definition is the same as the default but the volume of oil is at its bubble
point pressure rather than at stock tank conditions, that is:
i = std
1 std
R s ( p ) = -----------------------------
V o ( T, p bub ) ∑ Vg ( i ) [EQ 8.177]
i=p
The program option "Definition of Oil relative volume in Diff. Lib." on page 160 allows
an alternative definition of the oil relative volume where the volume of oil is
normalized to the initial volume of oil at reservoir rather than standard conditions:

Vo ( p )
B o ( p ) = ----------------------------
- [EQ 8.178]
V o ( T, p bub )
Gas condensate systems

Essentially three experiments are performed on gas condensate systems:
• "Dew point evaluation" on page 358.
• "Constant composition expansion" on page 358.
• "Constant volume depletion" on page 359.
Hint It is not uncommon for CCEs to be performed on all fluids, and it is

recommended to perform CVDs on more volatile oils.
Dew point evaluation

This experiment is generally more difficult than the bubble point evaluation and
consequently subject to larger uncertainties.
1 Most condensate systems are relatively compressible above the dew point so that
the appearance of a heavier, less-compressible second liquid phase cannot be
identified on a pressure/volume relationship. This means the dew point must be
found by eye, by visual determination of the pressure when the first drop of liquid
is formed in the cell.
Note This can be affected by imperfections or grease in the cell, poor experimental
procedures, etc., and it is not uncommon for errors of 100 psia to be associated
with this measurement.
PVTi
In PVTi, this experiment is simulated in much the same way as the bubble point
evaluation except that the trial second (liquid) phase is heavier than the original fluid.
1 The same stability test is done to find the highest two phase state and the same set
of equations and variables iterated to solution.
2 As an alternative to this high-pressure dew point, usually referred to as the
retrograde dew point, the low-pressure or normal dew point can be found by
starting low in pressure and iterating higher.
Constant composition expansion

This experiment is often done while trying to find the dew point of a gas (or bubble
point of a volatile or even crude oil).

1 It consists of varying the pressure and measuring the resulting volume of the single
phase fluid above saturation pressure, and volumes of vapor and liquid (and total)
phases below it.
2 For the single phase state, the vapor Z -factor or liquid density can be calculated
from the other fluid properties.
These data items are reported along with the relative volume, being the volume of the
fluid at any given pressure per the volume of the fluid at the saturation pressure.
PVTi
In PVTi, the CCE is simulated by:
1 Finding the saturation pressure which for one mole of feedstream defines the cell
or control volume.
2 Then at pressures above the saturation pressure it is sufficient to find the single
phase liquid or vapor Z -factor to calculate the volumetric behavior.
3 Below the saturation pressure, the feedstream is flashed at each required pressure
stage to determine the phase split and other volumetric properties.
Constant volume depletion

The CVD is the most useful, and probably the most difficult experiment to perform on
a gas condensate (and increasingly on volatile oils).
1 It consists of starting with a volume of fluid at its saturation pressure, which again
defines the cell or control volume for the experiment.
2 Then the pressure is dropped by several hundred psia, or so, whereupon the fluid
becomes two phase and expands in volume.
3 Any excess volume over and above the cell volume is removed by taking off gas
which is analysed compositionally and volumetrically as well as noting the
number of moles.
4 In addition, the percentage of liquid in the remaining fluid, the cell volume, defines
the liquid saturation.
Generally reported is:
• S liq : Liquid saturation;
• N pro : Moles of vapor removed;
• Z gas : Z -factor of removed vapor;

• y i, j : Composition of each removed vapor stream;
and occasionally:
• x i, N : Composition of liquid stream left in cell at last pressure stage;
vap Mole weight of removed vapor plus fraction;

• M N+ :
vap Specific gravity of removed vapor plus fraction.

• γ N+ :
PVTi
In PVTi this experiment is simulated by:
1 First finding the saturation pressure. The volume occupied by one mole of fluid at
its saturation pressure then defines the cell or control volume.
2 At some pressure less than the saturation pressure, the fluid is flashed into two
phases and any excess gas is removed to return the volume to the control volume.
3 The number of moles of vapor and its properties are noted.
4 The composition of the remaining fluid is calculated by volumetric balance and
this forms the feedstream for the next pressure stage, and so on.
Process simulation
For the purposes of process-type simulation and/or for the definition of blackoil tables
for reservoir modeling, separator tests are performed to see what phase splits are
achieved when a fluid is flashed at a series of pressures and temperatures in some
prescribed sequence.
1 Most laboratory analyses consist of the reservoir fluid being flashed in a cell at
some specified pressure and temperature, and (generally) the liquid output being
fed to a second (and possibly third) cell at some reduced pressure and temperature:
the last stage cell usually being at standard conditions.
Volumes of gas evolved from each stage are generally collected together and the
properties of the resulting mixture are quoted/evaluated at standard conditions.
PVTi
1 In PVTi this process is simulated by a set of flash calculations at the required set of
pressures and temperatures, taking the reservoir fluid as the feedstream and
routing the liquid and vapor outputs to other stages.
2 The default output streams are liquid to the next stage and vapor to the stock-tank
accumulation, but PVTi is capable of solving feedback loops, where, for example,
vapor output is routed back to an earlier stage.
3 In addition, a user option allows the output streams to be split and directed to
more than one other stage.

Optimized separators
In the Fluid Properties (FPE) workflow, PVTi offers the automatic generation of the
optimum separator configuration for a fluid. The optimum separator configuration is
defined as the two-stage separator for which the stock-tank formation volume factor
(Bo) is minimized.
Given the reservoir temperature and the maximum separator pressure, the following
separator chain is constructed:
1 A separator stage at some temperature below the reservoir temperature and some
pressure below the highest separator pressure. This stage is optimized.
2 A second stage at standard conditions, the liquid output from Stage 1 passes into
this second stage.
3 The stock tank. The vapor output from Stages 1 and 2 both pass into the vapor
stock tank and the liquid output from Stage 2 passes into the liquid stock tank.
A special regression procedure is used to determine the optimum temperature and
pressure for Stage 1, such that Bo is minimized in the stock tank. This point coincides
with minimum Total Gas-Oil Ratio and minimum stock tank vapor Gravity.
Note If no maximum pressure is supplied, it is assumed that the maximum

separator pressure available is 1440 psia (source: Oil Phase).
For more information on fluid properties estimation see "Fluid Properties Estimation"
on page 406.
Gas injection processes

PVTi has several simulations available for investigating gas injection processes.
The three that correspond closely to laboratory experiments are:
• "Swelling test" on page 361
• "Vaporization test" on page 362
• "Multiple contact test" on page 362
In addition, PVTi has available: a ternary diagram, and first and multiple contact
minimum miscibility pressure experiments by one-cell simulation.
Note These have no equivalent in the laboratory.
Swelling test
1 The swelling test consists of finding the saturation pressure and hence volume of
a reservoir fluid.
2 Followed by adding, in a series of steps, prescribed volumes (or moles) of lean
injection gas, re-pressuring the resulting mixture to return to a single phase system
and measuring the new saturation pressure and volume.

The data quoted is the set of saturation pressures for the original fluid and the mixtures
and the ratio of the saturation volume of the mixtures to the saturation volume of the
original fluid, usually referred to as the swelling factor.
PVTi
In PVTi this is simulated by:
1 Firstly finding the saturation pressure of one mole of the required reservoir fluid
and hence the saturation volume.
2 Then prescribed volumes of a lean injection gas are added as a GOR (volume of
injection gas at standard conditions per volume of original reservoir fluid at its
Psat or other prescribed pressure) or a mole% (moles of lean gas per moles of
mixture) to give a new fluid composition.
3 The saturation pressure and volume of the new mixture are found, and hence the
swelling factor.
Vaporization test
This is similar to the swelling test, except that it is performed at constant volume and
pressure.
1 A volume of reservoir fluid at some pressure and temperature, usually below the
saturation pressure and hence two phase, is contacted by a series of lean gas
injections.
2 After each contact, a volume of (enriched) gas and/or liquid is removed to return
the system to the original volume. The composition of the removed gas is
measured.
The experiment seeks to measure the extent of vaporization of intermediate and heavy
components from the reservoir liquid phase by stripping into the injected gas stream.
In PVTi, the procedure is as follows.
1 One mole of reservoir fluid at a prescribed pressure and temperature is flashed to
calculate the phase split and volumes of liquid and vapor.
2 Lean gas is then added in a series of steps of moles, to give a new mixture
composition which is flashed at the same pressure and temperature, and any
excess fluid volume removed.
3 The composition of the removed stream is noted and the resulting mixture is then
subject to further lean gas injections.
Multiple contact test

PVTi
1 In condensing drive mode the initial reservoir oil (the mole fraction can be
specified) is contacted with one mole of injection gas at a specified temperature
and pressure.
2 This mixture is then flashed and a specified fraction of the resulting oil is then
contacted with one mole of the initial injection gas. This process is repeated at each
stage of the test.

The vaporizing drive follows the same procedure as that of the condensing drive
except that it is a specified fraction of the resulting gas from each flash that is contacted
with one mole of reservoir oil.
A special ternary plot is available for this experiment that plots the compositions at
each stage of the oil and gas resulting from each flash, and these points effectively mark
the boundary of the two-phase region at the specified conditions.
Variation of composition with depth

PVTi
Although not a laboratory experiment, the estimation of the variation of a fluid’s
composition with depth is of great possible value.
1 On purely thermodynamic principles, fugacity in an isothermal system that can be
expected to vary with depth according to:
Mi g ( h – h0 )
ln f i = ln f i, 0 + -----------------------------
- [EQ 8.179]
RT 0
2 Assuming thermodynamic equilibrium in the fluid at some reference height h 0

(namely the equal fugacity conditions and mole composition constraints of
"Material balance" on page 330 and "Flash calculations" on page 330), the N -
equations, defined in [EQ 8.179], and a further mole constraint equation can be
used to determine the N -compositions and pressure, z i , p , from the reference
composition and pressure, z i, 0 , p 0 .
Note However, the compositional gradient experiment assumes many conditions:

namely thermal, gravitational and diffusive equilibrium. Any or all of these
conditions can be violated in a given reservoir, thus invalidating the use of [EQ
8.179]).
Note The existence in some reservoirs of a temperature gradient indicates a lack of

global thermodynamic equilibrium, but may still allow a steady-state
situation, in which the mass flux is zero (that is hydrostatic equilibrium),
whilst the energy flux is not. As yet, there is no consensus on how to treat such
temperature gradients, and thus PVTi does not attempt to model them at the
moment.
With the above caveats, the experiment may still be of use in the prediction of the
existence and possible location of either gas-oil contacts or so-called critical transitions.
This latter transition consists of the grading of a fluid from an oil to a gas, or vice-versa,
without passing through a gas-oil contact. This occurs when an over-pressured
reservoir has a fluid composition at some depth whose critical temperature is equal to
the reservoir temperature at that depth. There is evidence that such behavior exists in
some North Sea fields.
Critical point experiment

This experiment calculates the critical point of the given sample.

FCMP experiment
Refer to Jensen and Michelsen [Ref. 38] and Pedersen et al. [Ref. 30] for detail.
MCMP experiment
Refer to Jensen and Michelsen [Ref. 38] and Pedersen et al. [Ref. 30] for detail.
Tsat experiment
Refer to Michelsen [Ref. 15] for further details.

Regression
Introduction
• "Practical considerations" on page 365.
• "Theoretical model" on page 369.
Practical considerations
There are no set rules for how to do regression of an equation of state model to match
to laboratory measurements.
The paper by Coats and Smart, [Ref. 27], contains an appendix on the choice, selection
and range limits of regression variables.
However, the Coats and Smart model is limited in its choice of regression variables to
the Ω a ’s, Ω b ’s and the binary interaction coefficients.
Another limitation of their model is the use of the standard two-parameter equation of
state. It is well known that the two-parameter Equation of State is in error in prediction
of liquid properties by as much as 10%. Therefore, to get matches to saturation
pressures and densities it becomes necessary to change the properties of “well-
defined” components, say Ω a ( C 1 ) and Ω b ( C 1 ) .
Such a problem is avoided in PVTi by use of the PR3 or SRK3 Equation of State, which
allows the volume shift parameter to be a possible regression variable.
Additionally, rather than varying just the Ω a ’s and Ω b ’s, PVTi allows you to change
the critical pressures and temperatures. This has the advantage in that monotonicity
tests can readily be applied to the set of T c ’s, p c ’s,..., etc., for the hydrocarbon
components to ensure that critical temperatures increase with increasing mole weight,
critical pressures decrease (except C 1 , C 2 , C 3 ) etc. For further information on
performing a regression in PVTi see "Regression in PVTi" on page 137 and "REGRESS
section keywords" on page 170.
The present section also contains information on:
• "Consistency and quality of measured data" on page 365
• "Plus fraction" on page 366
• "Performing regressions in PVTi" on page 367.
Consistency and quality of measured data

As a first step, before any regression is considered, as many tests as possible should be
performed to test the consistency and quality of the measured data. Clearly it is not
possible to match to an inconsistent PVT report, yet our experience is that all reports
are flawed to a greater or lesser degree.
The cause of these errors may be numerous. For example:

Regression
• Poor sampling and/or collection
• Bad laboratory procedures
• Simple typing mistakes in reporting (for example a composition does not add up
to 100).
It is most likely to be a combination of all these effects. However, an equation of state
model cannot be used in isolation from measured data as no two fluids are ever likely
to be the same.
PVTi provides you with the ability to check the consistency of CVD reports in the COMB
section. For further information see "COMB - Compositional Material Balance" on page
123 and "COMB section keywords" on page 170. This test should always be done if the
data is available; the results may be surprising.
Simple tests like checking compositions sum to 100% often reveal errors.
Another useful check is to plot pressure-dependent data, for example liquid dropout
from a CVD, relative volumes from a CCE, etc., to see if they are smoothly varying. One
way this can be done is to enter this and other data as OBS to compare with the
Equation of State predicted values from the experiments in the SIMULATE section. For
further information see "Simulation using PVTi" on page 128 and "SIMULATE section
keywords" on page 170.
Generally, the properties of the multi-stage experiments vary smoothly (apart from
discontinuities in vapor/liquid properties across a saturation pressure boundary), so
that vapor Z -factors, etc., that decrease, increase and decrease again as pressure drops
in a CVD, probably indicates data error.
Check the definitions of measured data. It has been our experience that what one
laboratory may call liquid saturation in a CCE experiment, for example liquid volume
divided by volume of the fluid at the saturation pressure, may be different to that from
another laboratory, say liquid volume divided by the current cell volume.
Check that a consistent set of units is being used. The saturation density of a crude oil
3
is often quoted as a specific volume in units of ft /lb , for example.
Plus fraction
Having determined that the data is reliable, or having rejected either poor data or
adjusted it to be consistent, one will usually find that the Equation of State predictions
differ from the measured data.
Most of this error can probably be associated with the incomplete fluid description,
namely the failure to fully characterize the plus fraction, for example C 7+ , although
some error has to be associated with the inadequacies of a cubic Equation of State.
Clearly, the plus fraction, consisting as it does of many hundreds or possibly thousands
of components, cannot be represented by just one component without some
modification. Even then, just to represent a C 7+ of a condensate or volatile oil, which
may be 5-20 mole percent of the fluid, by a single component with regressed properties
may be insufficient for describing a multi-pressure process.
On some CVD analyses the mole weight and specific gravity of the produced gas plus
fraction are measured and reported.

Regression
In such a case you may note that the plus fraction gets lighter as one would expect.
Then it is clearly not adequate to represent the plus fraction by a single component and
one should consider splitting it into two or three pseudo-components, say using the
techniques available in PVTi.
The properties of the plus fraction, or its pseudo-components if splitting has been
performed are the obvious candidates for regression. However, one should avoid
regressing the property of a component with a small mole fraction, say 1%.
This can be avoided by using the group facility in regression in which an Equation of
State property of two or more components can be treated as a single variable. This
grouping of components for regression should be borne in mind if the aim of the PVTi
analysis is to produce a pseudoised set of components for a compositional simulation.
Finally, one should always use the minimum set of variables possible.
It is not possible to say what this set is for any given fluid/measurement set, but the
symptoms of redundant variables are easy to spot; for example, one or more of the
variables is hitting one of its limits, or there is bouncing of the reported Rms error
within a small range.
If the regression facility from an interactive session is being used, you have the option
(as default) of rejecting the latest regression and restoring the pre-regression Equation
of State.
This makes it possible to vary the set of variables and test the success, or otherwise, of
a given set of variables in a limited number of iteration steps, for example, 10. The
sensitivity of the observations to which one is attempting to match as a function of the
regression variable set is output both to the screen and to the PVP printable. This gives
a direct measure to the relative importance, or otherwise, of a given variable.
Performing regressions in PVTi

Whilst not being complete, the following may be of assistance:
1 Always use one of the three-parameter Equation of State. We suggest the Peng-
Robinson Equation of State (PR3). The extra degree of freedom allows the
possibility of matching saturation pressure using critical properties etc., and then
independently matching to saturation density (liquid) or Z -factor (gas) using the
volume shifts.
2 Consider splitting the plus fraction for volatile fluids, that is gas condensates and
volatile oils. Genuinely, crude oils are often well described by a single plus fraction,
such as the C 7+ reported. Dry gases do not have a significant quantity of plus
fraction to affect results. Of all the splitting algorithms currently available, the best
appears the modified Whitson method, otherwise referred to as Semi-Continuous-
Thermodynamics (SCT). The use of this model allows the use of the special
regression parameters: sample by sample plus fraction mole weight, probability
density function skewness parameter and overall PNA-distribution.
3 There is strong evidence to suggest that the adjustment of one or two binary
interaction coefficients to create an un-symmetric pattern is very dangerous, see
[Ref. 30]. The alternative is to select the Cheuh-Prausnitz BICS (see "Binary
interaction coefficients" on page 353) using the appropriate “options” switch, and
then regress the pre-multiplying A -coefficient to adjust the binaries.

Regression
4 As an alternative to 3., one might consider other properties. If the SCT-splitting has
been activated, one might consider the mole weight and skewness parameter or
the overall characterization of the plus fraction. These techniques have the
advantage of preserving monotonicity of all the key Equation of State variables,
critical temperatures and pressures, and acentric factors. If choosing the critical
properties or acentric factors, one should start with just those of the plus fraction,
or the pseudo-components split from it.
If this is not sufficient, remember that any SCN-group in your sample analysis (for
example C 6 , C 7 ,...) are subject to uncertainty because of their PNA-content. One
should not have to consider changing properties associated with C 1 , C 2 ,..., and so
on. However, if you have a significant mole percentage of inorganics in your fluid
or they are being injected into your fluid, it may be appropriate to adjust the set of
inorganic-hydrocarbon binaries, probably as a single group change.
5 Critical volumes or Z -factors (equivalent variables) are only needed for the
Lohrenz-Bray-Clark (LBC) viscosity correlation.
Note The LBC correlation can be as much as 150% in error and is regularly 50% in
error.
The Pedersen correlation appears to be much better (and is insensitive to V c , Z c ).

This having been said, the fact that Z c ’s or V c ’s affect only LBC viscosities means
they can be regressed independently of all other variables/results. If doing this, it
is suggested they be regressed simultaneously as a single group because of the LBC
functional form, [EQ 8.126].
PVTi now has independent sets of critical volumes and Z -factors, one set are used
in the LBC viscosity correlation and can be selected as regression variables, the
other for use in the equation of state. The latter are not available for regression as
they do not affect results.
6 For characterized components try using the molecular weight of these
components. This is available as a special variable and must not be used in
conjunction with Tcrits etc., as changing the molecular weight of each
characterized component changes most of the properties by way of the
characterization technique anyway. The use of this variable is a very powerful
alternative one to the approach of Tcrits, Pcrits and acentric factors etc. as described
above
All observations entered into PVTi can be assigned a weighting factor that multiplies
the measured minus calculated residuals, see [EQ 8.181].
Clearly, some observations are more important than others and should be given a
higher weight to account for this.
As a general rule, the saturation pressure should be given the highest weight followed
by saturation density and then other quantities.
Users who have performed several material balance calculations on CVD experiments
will probably recognize that mole compositions should generally be given low
weights, if they are to be used at all.
The only way to perform regression is by trial and error.

Regression
Define your experiments and associated (consistent and reliable) observations and
save them to a PVI file prior to starting regression.
Include different sets of variables, experiments and observations. Use engineering
judgement to decide when the best match has been achieved to the maximum amount
of data while remembering to maintain monotonicity and physical consistency in one’s
Equation of State model.
Theoretical model
Generally the results of an equation of state model must be tuned by regression of one
T
or more variables, x = ( x 1, …, x N ) to a set of laboratory experiments,
˜
T
y = ( y 1, …, y M ) [EQ 8.180]
˜
where M ≥ N and:
yi = yi ( x )
˜
or in residual form:
ri ( x ) = wi [ ( yi – yi ( x ) ) ⁄ yi ] [EQ 8.181]
˜ ˜
th
where w i is the weighting factor applied to the i item of observed (or measured)
data, y i .

• "Regression algorithm" on page 369.
• "First and second order derivatives" on page 370
• "Trust region" on page 370
• "Termination conditions" on page 370
Regression algorithm
The regression algorithm in PVTi seeks to minimize the least squares residual
( l 2 norm) given by the objective function:
M
1 2 1 T
f ( x ) = ---
˜ 2 ∑ r i ( x ) = --- R ( x ) R ( x )
˜ 2 ˜ ˜
[EQ 8.182]
i=1
where R ( x ) is the residual function.

˜
The minimum of f ( x ) occurs at some x∗ , where:
˜ ˜
∇f ( x∗ ) = 0 [EQ 8.183]
˜
Since f ( x ) is generally non-linear in the set x , [EQ 8.183] must be solved iteratively.
˜ ˜
Applying Newton’s method gives:

Regression
–1
x j + 1 = x j – ( ∇2f ( x j ) ) ∇f ( x j ) [EQ 8.184]
˜ ˜ ˜ ˜
First and second order derivatives

The first and second derivatives of the objective function can be readily evaluated, see
Dennis and Schnabel, [Ref. 9].
It is possible to construct a numerical approximation to the first derivative using two
function evaluations. That is with the current values of the variables, x , and shifted
˜
values x + δx .
˜ ˜
A similar procedure to evaluate the second derivative becomes prohibitively expensive
and so an approximation is generated, see [Ref. 9].
The advantage of this pseudo second-order method is that it ensures that the algorithm
proceeds towards a genuine minimum.
A function of a single variable, say g ( x ) , has a minimum at the point x m if g' ( x m ) = 0
and the second derivative g'' ( x m ) is positive.
For a multi-variate function like the objective function, f , the equivalent requirement
is that the matrix ∇2f ( x ) is positive definite.
˜
A matrix is positive definite if it is symmetric and all its eigenvalues are positive.
The approximation used to construct the second derivative ensures the symmetry
condition. The program can correct for the lack of positive eigenvalues by adding a
multiple of the identity matrix.
Trust region
The algorithm only approaches the true solution rapidly if the current estimate x j is
˜
close to the actual solution x∗ . Generally, the objective function at any iteration j is
˜
only an approximation to the true solution. Therefore, it helps to identify a region
surrounding the current solution, in which we trust our second order approximation
to model the actual second order problem.
PVTi employs the trust region model by defining a step length, δ c , which is a measure
of the region in which our model can be thought to be representative.
Ideally, the step-size, which is determined automatically subject to certain maximum
and minimum sizes, decreases as the iterations proceed to convergence.
The program uses the value of the current step length to adjust the second order
derivative matrix, so that a zero or small value of δ c causes PVTi to solve the full
Newton problem, refer to [EQ 8.184].
If following the evaluation of a step, the program determines that the objective
function would not decrease, then the algorithm has a procedure for cutting back the
step so that it ensures a decrease in f .
Termination conditions
A variety of termination conditions from the regression algorithm are possible.

Regression
Note The required condition is that the residual goes to some small value, such as
–6
10 , but this is rarely achieved for larger problems.
An alternative is that the gradient of the objective function goes to zero, indicating that
the regression is approaching a minimum and no further improvement can be
expected.
As a result of one or more steps made by the algorithm, one or more of the selected
regression parameters can be pushed out of physical range.
What constitutes a physical range is debatable, but it is suggested that an increase or
decrease in excess of 50% of the original value causes termination. You can change
these limits but care should be taken.
You can request a premature termination by allowing only a small number of
iterations, say five or ten.
Hint This is a good practice to follow for any regressions which involve several
variables, experiments and observations.
Finally, if any errors are detected in any of the experiments during the regression, then
depending on their nature, the program may terminate the regression.

Regression
Output for ECLIPSE simulators
General information
Our suite of reservoir simulation software supports five different models for fluid
behavior. There is the ECLIPSE extended blackoil model, the pseudo-compositional
model (GI option), the ECLIPSE compositional model., the ECLIPSE Thermal model
and the API Tracking option used in ECLIPSE BlackOil.
p - V - T data suitable for use in any of these models can be generated from the PVTi
package. Each of the models and typical output is discussed in the appropriate section.
• "Blackoil model" on page 372.
• "Differential and composite from differential tables" on page 379.
• "Pseudo-compositional tables for ECLIPSE GI option" on page 380.
• "Compositional data for ECLIPSE Compositional" on page 381.
• "Water properties" on page 382.
• "Model for API Tracking option in ECLIPSE BlackOil" on page 383.
• "Compositional Data for ECLIPSE Thermal" on page 387.
Blackoil model
ECLIPSE has a so-called extended blackoil model. That is, in addition to the standard
blackoil parameters, R s , B o and B g , it contains the parameter R v for modeling oil
vaporization in the gas stream.
Methods of obtaining black oil tables from compositional data are described by
Whitson and Torp, [Ref. 6] and Coats, [Ref. 3].
In both cases the basis of the method is a constant volume depletion experiment, used
to supply reservoir liquid and vapor compositions at a series of pressures. (In the case
of crude oil samples, a differential liberation process is used in the same way.)
The blackoil model, used by ECLIPSE, can be viewed as a two-component
compositional model. The “components” are stock tank oil and stock tank gas that are
assumed to be invariant and are assigned constant densities.
PVT model
The actual form of the PVT model for the hydrocarbons depends on whether:
• There is or could be liquid and/or vapor in the reservoir during its production.
• The reservoir phases produce stock tank oil and/or gas on flashing to surface
conditions.

Reservoir fluids
Assuming the stock tank fluids are constant, the reservoir fluids, named liquid and
vapor to distinguish them from the surface fluids which are designated as oil and gas,
are generally a combination of the stock tank fluids.
The amount of surface gas dissolved in the reservoir liquid is given by the Gas-Oil-
3 3
Ratio (GOR), denoted R s , which has units of sm /sm (metric), mscf/stb (field) or
3 3
scm /scm (lab).
The volume of surface oil vaporized in the reservoir vapor is given by the Condensate-
3 3
Gas-Ratio (CGR), denoted R v , which has units of sm /sm (metric), stb/mscf (field) or
3 3
scm /scm (lab).
The model uses formation factors to account for changes in volume when the fluids are
transferred from the reservoir to the surface. The factors for the liquid and vapor are
3 3 3 3
denoted, B o and B g , and given units of rm /sm (metric), rb/stb (field) or rm /sm
3 3 3 3
(lab), and rm /sm (metric), rb/mscf (field) or rcm /scm (lab), respectively.
The model assumes that the reservoir has been depleted to a pressure below the
saturation pressure (liquid bubble point or vapor dew point) and consequently has
become two phase.
One mole of reservoir vapor and liquid occupies the volumes V gr and V or ,
respectively. If each of these volumes are then flashed through some separator system
(two stages are indicated above but this can be variable) to stock tank conditions, then
most generally each reservoir phase partitions into the streams named stock tank gas
and oil.
Stock tank components

If the one mole of reservoir vapor, volume V gr , gives N gg moles of stock tank gas of
volume V gg and N og moles of stock tank oil of volume V og , whilst the one mole of
reservoir liquid, volume V or , gives N go moles of stock tank gas of volume V go and
N oo moles of stock tank oil of volume V oo , then conservation of mass requires that:
M gr = M og + M gg [EQ 8.185]
M or = M go + M oo [EQ 8.186]
where ( M gr, M or ) are the reservoir molar masses and ( M og, M gg ) , ( M go, M oo ) are the
corresponding actual surface masses after flashing.
Since ρ = M ⁄ V , we can write equations [EQ 8.185] and [EQ 8.186] as:
ρ gr V gr = ρ og V og + ρ gg V gg [EQ 8.187]
ρ or V or = ρ go V go + ρ oo V oo [EQ 8.188]
or:

1
ρ gr = ------ ( ρ gg + R v ρ og ) [EQ 8.189]
Bg
1
ρ or = ------ ( ρ oo + R s ρ go ) [EQ 8.190]
Bo
where:
V go V og
R s = --------
- R v = --------
- [EQ 8.191]
V oo V gg
V or V gr
B o = --------
- B g = --------
- [EQ 8.192]
V oo V gg
Note Strictly, the blackoil model requires that the stock tank “components” are
constant and invariant with time.
CVD process
In the Coats method, only the reservoir vapor is taken through the separator, the oil B o
and R s values being obtained by solving the mass conservation equations:
ρ sto V 2 ( b g S g + b o R s S o ) = ρ sto V 1 ( b g S g + b o R s S o ) [EQ 8.193]

2 1
ρ stg V 2 ( b g R v S g + b o S o ) = ρ stg V 1 ( B g R v S g + b o S o ) [EQ 8.194]

2 1
for each expansion step of the CVD process from V 1 to V 2 , where ρ sto and ρ stg are
fixed surface densities, and b o = 1 ⁄ B o , b g = 1 ⁄ B g .
The stock tank densities are obtained from the output of the separators at the saturation
pressure. Solving these equations yields values at all pressures except the saturation
pressure, at which the Whitson method can be used.
Alternatively, a small initial pressure step from the saturation pressure may be
specified.
Note The constant volume expansion method is usually applied to condensates, but
can also be used for volatile oils.
The differential liberation approach is only suitable for oils.
In this case, the analogue of the Whitson method runs both reservoir oil and vapor
through the separators as before.
The analogue of the Coats method uses the conservation equations, [EQ 8.193] & [EQ
8.194] (although in this case S g1 is zero for each step), ρ sto and ρ stg are obtained from
the reservoir oil at the saturation pressure.

ECLIPSE 100 tables
For both processes, you can produce the ECLIPSE live oil, wet gas, dead oil and dry gas
tables. The respective keywords are PVTO, PVTG, PVDO, and PVDG.
In addition, the model calculates the density of the oil, water and gas phases at stock
tank conditions (final stage separator) for the saturation point fluid, and output under
the DENSITY keyword. These values are important as the reservoir properties are
calculated using the R s , R v , B o , B g from the stock tank volumes and densities.
Hint The "Choosing the unit type for PVTi" on page 155 program option gives you
a choice of outputting and plotting all the data (saturated and undersaturated)
or just the saturated data.
An example of a typical set of blackoil data (for a volatile oil fluid) is shown below:
-- Blackoil tables for sample ZI at T = 176.00000 deg F

-- With Peng-Robinson (three-parameter) EoS
-- And Lohrenz-Bray-Clark viscosity correlation
-- ECLIPSE 100 DENSITY data
-- Surface densities of oil, water and gas:
-- Units of FIELD
DENSITY
50.74699 62.427970.05842 /
-- Two stage separator at
-- Pressures 214.69590 14.69590
-- Temperatures 100.00010 60.00001
-- Pressures in PSIA Temperatures in deg F
--ECLIPSE 100 PVTG data
--(Constant volume depletion)
--Units are FIELD
--Method used: Whitson and Torp

PVTG
-- PRES RV BG VISC
-- PSIA STB/MSCFRB/MSCF CPOISE
1114.695900.000992.43709 0.0139554 --Saturated
0.00000 2.439190.0139489 / --Dry gas
1814.695900.002761.43004 0.0166381 --Saturated
0.00099 1.433500.0165733 --Undersat gas
0.00000 1.435430.0165380 / --Dry gas
2514.695900.007361.02359 0.0206166 --Saturated
0.00276 1.029290.0202678 --Undersat gas
0.00099 1.031500.0201382 --Undersat gas
0.00000 1.032720.0200669 / --Dry gas
3214.695900.016950.82309 0.0257392 --Saturated
0.00736 0.828510.0246867 --Undersat gas
0.00276 0.831110.0242117 --Undersat gas
0.00099 0.832110.0240316 --Undersat gas
0.00000 0.832670.0239322 / --Dry gas
3914.695900.034920.71605 0.0321610 --Saturated
0.01695 0.717250.0297284 --Undersat gas
0.00736 0.717890.0285431 --Undersat gas
0.00276 0.718200.0279877 --Undersat gas
0.00099 0.718310.0277758 --Undersat gas
0.00000 0.718380.0276584 / --Dry gas
4077.568010.041080.69977 0.0339734 --Psat
0.03492 0.699600.0329685 --Undersat gas
0.01695 0.699100.0306120 --Undersat gas
0.00736 0.698840.0293996 --Undersat gas
0.00276 0.698710.0288303 --Undersat gas
0.00099 0.698670.0286129 --Undersat gas
0.00000 0.698640.0284924 / --Dry gas
4614.69590 0.041090.65591 0.0368566 --Generated
0.03492 0.654470.0357868 --Undersat gas
0.01695 0.650290.0332692 --Undersat gas
0.00736 0.648060.0319664 --Undersat gas
0.00276 0.646990.0313522 --Undersat gas
0.00099 0.646580.0311171 --Undersat gas
0.00000 0.646350.0309867 / --Dry gas
--ECLIPSE 100 PVTO data
--(Constant volume depletion)
--Units are FIELD
--Method used: Whitson and Torp
PVTO

-- RS PRES BO VISC
-- MSCF/STB PSIA RB/STB CPOISE
0.49137 114.6959 0.34107 0.30770 --Saturated
1814.69590 1.32574 0.33605
2514.69590 1.31260 0.36369
3214.69590 1.30116 0.39067
3914.69590 1.29106 0.41705
4077.56801 1.28887 0.42310
4614.69590 1.28206 0.44285 /
0.79978 1814.69590 1.49404 0.23723 --Saturated
2514.69590 1.47369 0.26054
3214.69590 1.45641 0.28344
3914.69590 1.44147 0.30594
4077.56801 1.43827 0.31112
4614.69590 1.42836 0.32808 /
1.16269 2514.69590 1.67183 0.18640 --Saturated
3214.69590 1.64551 0.20520
3914.69590 1.62331 0.22377
4077.56801 1.61862 0.22806
4614.69590 1.60421 0.24214 /
1.63520 3214.69590 1.90440 0.14692 --Saturated
3914.69590 1.87000 0.16174
4077.56801 1.86286 0.16517
4614.69590 1.84116 0.17646 /
2.34029 3914.69590 2.25803 0.11383 --Saturated
4077.56801 2.24611 0.11645
4614.69590 2.21050 0.12507 /
2.56913 077.56801 2.37483 0.10661 --Psat
4614.69590 2.33394 0.11462 /
2.66303 4614.69590 2.37492 0.10622 --Generated
5151.82379 2.33402 0.11423 /
/
Both the PVTO and PVTG tables are extended to include properties of the
undersaturated oil and gas, respectively.
In the case of the PVTG table the model computes, at all values of R v between the
saturated value and the dry gas value ( R v = 0.0 ), a value for B g by adding sufficient
stock tank oil to the stock tank gas (both from the flash of the saturated gas to stock
tank) to give the required R v and then taking the ratio of this fluid's volume at
reservoir and stock tank volumes.

In the case of the PVTO table, the model compresses the saturated fluid from a given R s
node at all pressures from the saturated pressure up to the highest pressure. The ratio
of these volumes to the resulting stock tank oil volume (from flashing the saturated oil)
gives the undersaturated B o values.
Monotonicity and compressibility tests

If you specify a highest pressure that is higher than the saturation pressure fluid, then
PVTi generates “saturated” values of R s , R v , B o , B g , which satisfy all ECLIPSE’s
monotonicity and compressibility tests.
The value of B g continues to decrease monotonically with increasing pressure,
regardless of whether the pressure is above saturation pressure or not.
Additionally, for an oil where strictly no longer has any free gas, the model can carry
on compressing the bubble point gas to generate B g . It then requires that the R v is
weakly monotonically increasing at p > p sat , that is:
R v ( p ) = R v ( p sat ) [ 1 + ε v ( p – p sat ) ] [EQ 8.195]
–5
where ε v is small, say ∼ 1 ×10 .
The model then generates extrapolated oil properties, R s , B o to satisfy the ECLIPSE
compressibility test:
C tot = S o C o + S g C g [EQ 8.196]
where ( S o, S g ) are the saturations, and ( C o, C g ) are the compressibilities, of oil and
gas respectively, the latter given by:
1 – dB dR s ( B g – R v B o )
C o = ------ ------------o + --------- ------------------------------ [EQ 8.197]
B o dp dp ( 1 – R s R v )
1 – dB dR v ( B o – R s B g )
C g = ------ ------------g + --------- ------------------------------ [EQ 8.198]
B g dp dp ( 1 – R s R v )
Alternatively if you do not want these generated points, PVTi allows you to increase
the saturated pressure of the sample to above the maximum pressure used in the
depletion experiment by mixing the sample with a lighter gas sample.
The gas sample used is the vapor split-off obtained from performing a p sat calculation
on the original sample.
The model mixes this vapor with the sample and performs a new p sat calculation on
the new sample.
It then mixes the resulting vapor split off with the new sample and repeats the process
until a p sat forms that is higher than the maximum pressure stage. If this process is
successful you can save the final sample and use it in the calculation of the black oil
tables.

Note Oil and gas properties below the saturation point generated as described
above are also subjected to the ECLIPSE total compressibility test.
Note If negative compressibilities are detected, PVTi (with confirmation) solves for
intermediate points, which provide the necessary resolution to pass the
compressibility test.
Differential and composite from differential tables

As an alternative to the composite tables described above, you can produce a
differential black oil table.
This technique involves simulating a differential liberation experiment, and the values
of B o and R s are taken to be the normal definitions of oil formation volume factor and
gas-oil ratio, respectively, as defined in this experiment.
Note These definitions are detailed in the manual (refer to "Basic laboratory
experiments" on page 355 and "Gas condensate systems" on page 358, and to
D
equations [EQ 8.171] and [EQ 8.173]). The quantities are known here as B o
D
and R s .
From these differential quantities, it is possible to estimate composite values of B o and

R s using a number of formulae.
Two of the more popular methods are due to Moses, [Ref. 41], and Fragor, [Ref. 40]
(both now termed CM and CF respectively).
Both methods start by estimating a value for B o and R s at the saturated pressure ( p sat )
of the initial oil. This oil is passed through a separator chain and the values of
CF CM
B o ( p sat ) and B o ( p sat ) are given by the oil formation volume factor from the
CF CM
separator. Likewise the values of R s ( p sat ) and R s ( p sat ) are given the value of the
gas-oil ratio from the separator.
Once these two values are computed, the rest of the values from the lowest pressure up
to p sat are found from the following formulae.
Firstly the Fragor definitions:

CF D
CF ( B o ( p sat ) – 1 ) ( B o ( p ) – 1 )
B o ( p ) = 1 + --------------------------------------------------------------------- [EQ 8.199]
D
( B o ( p sat ) – 1 )
and

CF D
CF R s ( p sat )R s ( p )
R s ( p ) = ----------------------------------------- [EQ 8.200]
D
R s ( p sat )
The Moses expressions are given by:

CM D
CM B o ( p sat )B o ( p )
Bo ( p ) = ------------------------------------------
- [EQ 8.201]
D
B o ( p sat )
and
⎛ B CM ( p )⎞
R s ( p ) = R s ( p sat ) – ( R s ( p sat ) – R s ( p ) ) ⎜ --------------------------⎟
CM CM D D o sat
[EQ 8.202]
⎜ D ⎟
⎝ B o ( p sat ) ⎠
PVTi can provide all four types of black oil output, that is, the usual composite, the
differential, and the two composite from differential tables.
Note The output form for the differential and composite from differential tables is
by way of the PVTO/PVDO and PVDG keywords only.
Pseudo-compositional tables for ECLIPSE

GI option
At pressures less than the saturation pressure, p sat , compositional effects become
important.
Limitations of the ECLIPSE blackoil model

1 The blackoil approximation, discussed in the previous section, models
compositional changes by making the parameters, R s , B o , R v , and B g all
functions of pressure, which it determines from the reservoir and surface
properties of the liquid and vapor phases.
2 The extended blackoil treatment, that is variable R v as available in ECLIPSE,
cannot model gas injection into condensates at pressures less than p sat , without
careful consideration (see ECLIPSE keyword VAPPARS). Lean gas injected into a
saturated reservoir fluid causes the stripping of the light and intermediate
components from the reservoir fluid, resulting in an enriched gas phase and a
depleted liquid phase.
3 To model this process accurately requires a detailed fluid description and the use
of a fully compositional simulator using a many-component model of the fluid.
However, such is generally impractical, especially using a fully implicit
formulation since the number of equations, which need to be solved increases with
the product of the number of components and the number of gridblocks.

ECLIPSE GI option
One possible solution to this problem is to extend the three-parameter blackoil model
as used in ECLIPSE BlackOil, that is p , S w and S o , by adding a fourth
parameter/equation.
In the ECLIPSE pseudo-compositional model, this fourth parameter, denoted GI, is the
cumulative gas volume, which has passed over the volume of reservoir liquid in a
gridblock. As such, GI is equivalent to a gas-oil ratio of volume of injection gas at
standard conditions to volume of oil at reservoir conditions:
gas
V stc
GI = ----------
- [EQ 8.203]
oil
V res
oil
This model determines the oil volume at reservoir conditions, V res , for the first stage
of gas addition, from the GI=0.0, that is the standard blackoil model. It then calculates
gas
the gas volume at standard conditions, V stc , from the gas law:
add
gas n ZRT stc
V stc = ---------------------------- [EQ 8.204]
p stc
add
where n is the number of moles of gas added. By convention, at standard pressure
and temperature, p stc = 14.7 psia, T stc = 60.0 °F , and the gas-compressibility factor,
Z = 1.
Having defined a Constant Volume Depletion (CVD) experiment and a separator
network with the last stage corresponding to stock tank conditions, the model then
performs a calculation of the conventional extended blackoil tables. This defines the
GI=0.0 data.
It then adds lean gas to the reservoir fluid in a series of stages and generates the
blackoil table with the mixture. The ratio of the non-zero GI blackoil properties and the
GI=0.0 properties defines a set of GI-multipliers. This gives a two-dimensional set of
tables in ( p , GI) to describe the fluid behavior.
Detailed evaluations of the model’s performance have been undertaken and with
favourable conditions and tuning against a full compositional treatment, this model
can give reliable predictions.
Compositional data for ECLIPSE Compositional

You can run ECLIPSE Compositional in one of three modes:
• Blackoil
As ECLIPSE 100
• K -value
Specify K -values at given pressures

• Compositional
Equation of state.
It is envisaged that generally you will want to run ECLIPSE Compositional in
compositional (Equation of State) mode.
However, in blackoil mode, ECLIPSE 300 reads the same PVTO and PVTG tables as used
in ECLIPSE BlackOil.
In K -value mode, you must specify a set of K -values for each component at a set of
pressures spanning the operating pressure region. You can generate these, as in PVTi,
by performing a CVD experiment on the reservoir fluid.
Compositional mode
The Equation of State mode is essentially the same as that employed by PVTi for its
flash, saturation pressure, etc., calculations.
You must specify the number of components, the Equation of State required (the same
set as available in PVTi), and the critical properties, acentric factors, binary interaction
coefficients, compositions, etc.
Clearly, the more components used, the more accurate the simulation (assuming the
component set has been optimised to measured data); however, such simulations
require more CPU time.
Hint If running ECLIPSE Compositional in fully implicit mode, more than six or
seven components may become prohibitive for all but the largest of
supercomputers or workstations, in which case you should consider
pseudoisation of this fluid system to fewer components.
Water properties
ECLIPSE and VFPi consider water to be non-volatile (only exists as liquid water) and
immiscible with the hydrocarbon phases (water cannot dissolve in the hydrocarbons
or vice-versa). This simplifies the PVT treatment of water.
If water is present, the variation of water volume in the reservoir with pressure, V w ( p ) ,
is defined with respect to the volume of water at surface conditions, V stw , by the
formation volume factor, B w , where:
Vw ( p )
B w ( p ) = --------------
- [EQ 8.205]
V stw
3 3 3 3
which has units of rm /sm (metric), rb/stb (field) or rcm /scm (lab).
Note Note that VFP only supports metric and field units.
Within the models, B w is defined in terms of the B w at some reference pressure, p ref :

B w ( p ref )
B w ( p ) = ------------------------------ [EQ 8.206]
2
(1 + X + X )
where
X = C w ( p – p ref ) [EQ 8.207]
and C w is the water compressibility which is defined as:
1 dB w
C w = – ------- ⎛ ----------⎞ p = p [EQ 8.208]
B w ⎝ dp ⎠ ref
–1 –1 –1
that has units of barsa (metric), psia (field) or atmos (lab).
Water viscosity is modeled in the same way as the formation volume factor:
μ w ( p ref )
μ w ( p ) = -----------------------------
- [EQ 8.209]
2
(1 + Y + Y )
where
Y = C μw ( p – p ref ) [EQ 8.210]
and C μw is the water viscosibility which is defined as:
1 dμ w
C μw = – ------- ⎛ ----------⎞ p = p [EQ 8.211]
μ w ⎝ dp ⎠ ref
–1 –1 –1
that also has units of barsa (metric), psia (field) or atmos (lab).
An additional complication can arise if the effect of salt (brine) concentration is to be
modeled (ECLIPSE BlackOil and GI option only). See the ECLIPSE keyword
PVTWSALT for further details in this case.
Note Correlations for generating default values for all the above correlations can be
found in [Ref. 33].
Model for API Tracking option in ECLIPSE BlackOil

Introduction
The API Tracking facility enables ECLIPSE BlackOil to model the mixing of different
types of oil, having different surface densities and PVT properties.
Without the API Tracking facility, the presence of different types of oil in the reservoir
could be handled with the aid of PVT region numbers. Oil in PVT region 1 would have
its properties determined from PVT table number 1, and so on. However, this method
cannot model the mixing of oil types. Oil flowing from region 1 into region 2 would
appear to take on the properties associated with region 2.

The API Tracking facility essentially replaces the concept of PVT regions for oil. The
PVT tables used for determining the oil properties are selected at each time step
according to the average API of the oil in each grid block (or to be more precise, its
average surface density). A mass conservation equation is solved at the end of each
time step to update the oil surface density in each grid block, to model the mixing of
the different oil types.
When using ECLIPSE BlackOil, the API Tracking facility is turned on by the keyword
API in the RUNSPEC section. In dead oil systems the keyword RSCONSTT cannot be
used with API tracking, but RSCONST (defining a uniform R s value over the whole
field) can be used.
Using PVTi to Export the Tables

PVTi can be used to export the black oil tables in the required form outlined in the next
few sections. Simply open the Export Panel for API Tracking option panel by using File
| Export | API Tracking option in ECLIPSE BlackOil....
The user needs to select the samples for which the export is required. This is done by
selecting a set of samples and dropping them in the use box. Next, the required
keywords to export need to be specified in the normal way.
Hint Since the API Tracking functionality in ECLIPSE models the mixing of live oil
properties the usual procedure in PVTi is to use the API Tracking export
facility to write out a series of tables containing the PVTO (live oil) and PVDG
(dead gas).
You can then specify whether they want a gas table for each sample. Normally the API
Tracking option in ECLIPSE only uses a different oil table for each sample and just a
single gas table is used, as explained in the next section, and so this is the default. If this
default option is used then PVTi exports a single gas table for the sample with the
median gas density at surface conditions.
The separator and units to use for the export can be specified in the normal way. You
can also specify a few other straightforward options such as whether to write to full
double precision and whether to plot the tables.
When the export is performed PVTi orders the tables so that the oil surface densities
increase monotonically with table number, which is a requirement of ECLIPSE. Each
table also has comments associated with it specifying which sample the table is linked
to and what the oil surface density is for this particular table.
ECLIPSE requires that the Rs ranges of any live oil tables are the same. To ensure this
is the case PVTi analyzes all the samples you selected and calculates the Rs at the
maximum pressure in the DL experiment. PVTi then finds the sample with the
maximum Rs and then uses linear extrapolation to extend the black oil tables of the
other samples so that the Rs ranges for all the samples are the same.
See the next few sections for technical details of ECLIPSE’s API Tracking option.

The PVT properties
Two or more sets of PVT tables should be supplied, each set being associated with a
particular value of the API. The oil property tables are entered in the usual way (with
keyword PVTO or PVCO for live oil problems, or keyword PVDO for dead oil problems),
but with the requirement that all the PVTO or PVCO tables must have the same upper
and lower Rs values. From these tables, ECLIPSE constructs a set of internal tables that
have a common set of Rs nodes, which includes the Rs nodes of all the input tables. This
allows ECLIPSE to use an efficient method of interpolating between tables, for
intermediate values of the API. The restriction that the upper and lower Rs values must
be the same for each table prevents extrapolation of the input data during this process,
which may cause unphysical behavior.
The API values associated with each PVT table are set using the keyword GRAVITY,
which inputs the oil, water and gas gravities for each PVT table number. The API
values are converted internally to oil surface densities, using the formula given in
"Conversion factors" on page 970. The oil API values must decrease monotonically
with the table number. Alternatively the surface densities associated with each table
number can be input directly using the keyword DENSITY. In this case the oil surface
densities must increase monotonically with the table number.
The PVT properties of water and gas are not affected by the API Tracking option. The
same number of tables must be entered in all the PVT data keywords and the ROCK
keyword, but normally only the first table is actually used in all except the oil PVT data.
The unused PVT and rock compressibility tables can be defaulted by typing a single
slash (/) on a new line for each unused table in the keyword.
Grouping tables into PVT regions

The API Tracking facility over-rides the concept of PVT table regions for oil. The tables
used to determine the oil PVT properties are selected according to the surface density
of the oil in the grid block, instead of the block’s PVT region number. Thus the PVTNUM
keyword in the REGIONS section is not normally required with the API tracking
facility.
If however the concept of PVT regions for different types of oil is still required in an
API Tracking run, the oil PVT tables can be partitioned into groups for use in different
regions of the reservoir. The keyword APIGROUP must be entered, to specify the
maximum number of groups of oil PVT tables. The grouping of the oil PVT tables is
then determined by the order in which they are entered. Within each group, the oil API
gravities should decrease (or the oil surface densities should increase) monotonically
with the table number. A break in the required monotonicity is taken to indicate that
the subsequent tables belong to the next API group. An error is flagged if the total
number of API groups exceeds the maximum specified in the APIGROUP keyword.
The separate groups of oil PVT tables can then be used in different areas of the
reservoir. The REGIONS section keyword PVTNUM is used to specify which API group
is to be used for each grid block. For example, a cell in PVTNUM region 2 will use the
second API group to obtain the oil properties.
For water and gas PVT properties, the PVTNUM regions correspond to the actual table
numbers, as in runs that do not use API tracking. So a cell in PVTNUM region 2 still uses
table number 2 for its water and gas properties.

Setting the initial conditions
The initial API distribution throughout the reservoir is specified in the SOLUTION
section. If the initial conditions are determined by equilibration, the API can vary with
depth independently in each equilibration region. Keyword APIVD supplies a table of
API values versus depth for each equilibration region.
When setting the initial conditions by enumeration (keywords PRESSURE, SWAT etc.),
the keyword OILAPI is used to supply the initial API values for each grid block.
The API tracking calculation

The initial API values in the grid blocks are immediately converted into oil surface
density. The hydrostatic pressure gradient calculation takes account of the varying
surface density of the oil. The variation of other PVT properties with surface density
(bubble point pressure, formation volume factor and viscosity) is handled by
interpolating between PVT tables. The two adjacent PVT tables whose oil surface
densities straddle the oil surface density value in the grid block are located, and their
properties ( 1 ⁄ B o , 1 ⁄ ( B o μ o ) ) are interpolated linearly in oil surface density.
To calculate the bubble point, the two PVT tables corresponding to oil surface gravities
either side of the current grid block API value are located. The bubble point is then
obtained by linearly interpolating between the bubble points in these PVT tables at the
given grid block dissolved gas-oil ratio.
The oil surface density in each grid block and well is held constant over the time step.
When a converged solution for the time step has been found, and the inter-block flows
determined, a mass conservation equation is solved to update the oil surface densities.
The new densities are used in the next time step.
The oil surface densities are converted back into API values for the output reports. The
mnemonic OILAPI in the RPTSCHED keyword provides an output of the grid block
API values, and the well reports include the API values in the wells.
API tracking in wells

The oil surface density in each well is calculated at the end of the time step to reflect a
flow-weighted average surface density of the inflowing oil. A crossflowing well will
reinject oil of this average surface density back into the formation. However, the
hydrostatic head calculation in the wellbore will be based on a (more accurate) flow-
weighted average of upstream connection inflows if WELSPECS item 12 is ‘SEG’,
allowing the oil density to vary with position in the wellbore. (The pre-98A treatment
of wellbore hydrostatic head, which uses the well’s average oil surface density, can be
restored if required by setting switch 35 in the OPTIONS keyword.)
If the well uses a VFP table to calculate the tubing pressure losses, this should be
calculated using a suitable value of the oil surface density. However, it is possible to
take account of the variation of the well’s producing API over time by defining the fifth
lookup variable, the ALQ, to represent the surface density of the produced oil (see
keyword WALQCALC). The VFP table should be prepared using the same definition of
the ALQ, with values spanning the expected range of oil surface densities. The VFPi
program contains an option to prepare tables with variable oil surface density.

Behavior in wet gas systems
By default, in wet gas systems, the surface density property (API) of the oil is
transported in both the oil and gas phases, that is the API is interpreted as a property
of the oil component rather than the liquid oil phase.
By using OPTIONS switch 58 this behavior can be modified. When the switch is set
greater than zero the API property will only be transported in the liquid oil phase. In
this case the API property can be thought of as a heavy component which does not
vaporize. Care needs to be taken when using this option as it is possible to vaporize all
the oil in a cell, leading to high surface densities and negative API values.
Compositional Data for ECLIPSE Thermal

Introduction
This introduction contains a brief outline of the reason for the introduction of this
module.
• "Outline of keywords for ECLIPSE Thermal" on page 387 contains a brief
description of the keywords exported.
• "Workflow" on page 388 section gives step-by-step instructions of how to perform
a typical export for use in ECLIPSE Thermal.
• "Some handy hints for fitting the Crookston coefficients" on page 390 gives some
hints and tips on how to export Crookston coefficients that provide accurate
approximations to PVTi’s EoS K-values.
For the 2003A version of PVTi a new ECLIPSE Thermal support module was available
where you were able to interactively develop a correlation which accurately predicted
K-values for each component in a given fluid. For the 2004A version this module has
been extended to a full export facility where you can write out files that are suitable for
use as PVT input for ECLIPSE Thermal. The motivation behind this is so that, just as
you can export files to use as PVT input for ECLIPSE BlackOil and ECLIPSE
Compositional, they will now be able to do the same for ECLIPSE Thermal.
When you use this new export facility, they now have a choice of keywords to export
regarding K-values - namely the KVCR keyword (which was available in 2003A PVTi)
or the KVWI keyword which is tells ECLIPSE Thermal to use a simplified version of
Crookston’s equation called Wilson’s formula. In addition PVTi also exports a series of
other keywords enabling ECLIPSE Thermal to calculate properties of the feed fluid
such as oil density, gas density and oil/gas phase enthalpies.
For a technical description of how these keywords exported by PVTi are used in
ECLIPSE Thermal see "ECLIPSE Thermal Export Module" on page 423. For a summary
and brief outline of the nature of these keywords see the next section.
Outline of keywords for ECLIPSE Thermal

1 STCOND - the standard temperature and pressure used in PVTi
2 NCOMPS - the number of components in the fluid.
3 CNAMES - the names of each component.
4 MW - the molecular weight of each component.

5 TCRIT - the critical temperature of each component.
6 PCRIT - the critical pressure of each component.
7 TREFT - the ECLIPSE Thermal reference temperature at which the reference
density, DREFT, is calculated.
Note In general the quantities TREFT, PREFT, DREFT are not the same as the related
quantities of reference temperature, TREF and references density, DREF used
in PVTi. Please see
8 ACF - the acentric factors of the components.

9 SPECHA - oil phase specific heat first coefficients.
10 SPECHB - oil phase specific heat second coefficients.
11 SPECHG - gas phase specific heat first coefficients.
12 SPECHH - gas phase specific heat second coefficients.
13 HEATVAPS - heats of vaporization at the standard temperature.
14 CREF - component isothermal compressibilities in the oil phase.
15 THERMEX1 - component thermal expansion coefficients in the oil phase.
16 ZFACTOR - the gas Z factor for each component.
17 KVWI - tells ECLIPSE Thermal to use Wilson’s formula to estimate K-values.
18 KVCR - tells ECLIPSE Thermal to use Crookston’s equation to estimate K-values.
This keyword contains the values of the coefficients (A-E) for each component.
19 ZI - the feed mole fractions of each component
The keywords STCOND, SPECHA, SPECHB, SPECHG, SPECHH, HEATVAPS, CREF,
THERMEX1, ZFACTOR, KVWI and KVCR are only exported for ECLIPSE Thermal and not
for ECLIPSE Compositional. See "ECLIPSE Thermal Export Module" on page 423 for a
technical description of how the values contained in these keywords are used to
calculate K-values, oil/gas densities and oil/gas enthalpies for each component.
Workflow
1 To access the Thermal Support module select Edit | Samples | ECLIPSE Thermal
Support | Export for ECLIPSE Thermal.... Alternatively right- click on a fluid
sample of interest and select Export ECLIPSE Thermal Model....
2 In the panel type in the maximum and minimum values of the pressure in your
reservoir and also the maximum and minimum temperature.
3 Select the sample you wish to export the model for by using the drop-down menu.
4 You need to decide if you want to estimate K-values in ECLIPSE Thermal using
Crookston’s equation or using Wilson’s formula. Wilson’s formula gives much less
accurate results than Crookston’s formula but if you have time constraints then
much less work is involved in exporting this information from PVTi. If you want
to use Wilson’s formula then your work is finished and you do not need to go
though steps 5-12. Simply select which units you want to export and click OK.

Hint If you have time and you know the rough limits on temperature and pressure
within your reservoir then Crookston’s equation will, in general, give you
much more accurate estimates of K-values than Wilson’s formula.
5 If you want to export coefficients for Crookston’s equation then tick the box
entitled Export Crookston Coefficients?
6 Now enter how many flashes you want to perform.
Unless you have very large ranges in temperature or pressure in your reservoir
then the default of 20 is probably fine. These flashes generate K-values using
PVTi’s EoS algorithm at random points in the region defined by your maximum
pressure and temperature. The generated K-values are the “observations” used
when trying to optimize the values of the coefficients in Crookston’s equation.
7 Click the Fit Crookston Coefficients button.
Another panel opens.
8 Select the coefficients to optimize. As a rule always start with B and D or A,B and
D. Now click Apply and PVTi attempts to use a minimization algorithm to calculate
the best values of your chosen coefficients that gives the most accurate prediction
of K-values for each component.
9 Once PVTi has finished a results window opens with two folders. The Coefficients
folder enables you to see the coefficient values derived by PVTi for each
component. The Statistics folder enables you to observe the mean rms of the fit (in
%) and the standard deviation (in %) of the fits for each component over the
specified temperature and pressure range.
10 To view your fit click View Fit on the Fit Crookston Coefficients panel. Another
panel opens. You are able to view the results by plotting K-values as a function of
temperature (at constant pressure) or pressure (at constant temperature). Type in
the appropriate information and click Apply.
11 PVTi now show you the observations it was fitting to for each component, that is
the K-values found using the Equation of State based flash (the points) and the
curves, which are the predictions of the K-values found using Crookston’s
formula.
12 Once you are happy with your results close the Fit Crookston Coefficients panel
and click OK on the Export for ECLIPSE Thermal panel. A File | Save panel opens.
Choose the name of the file you want to save it as and click Enter. A text file is
created and shown in the output display containing the exported keywords. This
file can be used as the PVT input in an ECLIPSE Thermal simulation.
Note The default file extension when exporting for ECLIPSE Thermal is .PVOT. The
save panel assumes you want to call your exported file filename.PVOT
where filename.PVI is the name of your PVTi project file.
See the next section for some hints on performing the optimization of the Crookston
coefficients.

Some handy hints for fitting the Crookston coefficients
The goal of this optimization is to allow you to interactively develop a correlation that
accurately predicts K-values for each component in a given fluid over a specified range
of temperature and pressures. The functional form of the correlation used by PVTi is
given by Crookston’s equation which is:
K ( P, T ) = ⎛⎝ A + --- + CP⎞⎠ exp ⎛⎝ -------------⎞⎠

B –D
[EQ 8.1]
P T –E
You are able to specify which coefficients to use in equation [EQ 8.109] (as already
explained in the previous section) and PVTi calculates the values of these coefficients
which give the best fit to the PVTi flash for each component over the pressure and
temperature range. The following things may help you in your workflow:
1 Try doing a phase curve first of your fluid to get an idea of the pressure and
temperature range where your sample is two-phase. PVTi does not use any regions
of pressure-temperature space where the fluid is single phase.
2 When choosing the values of the max/min temperature and pressure consider the
following:
• The default values of Pmax=1000psia, Pmin=50psia, Tmax=400F and Tmin=50F
are considered to be reasonable variations of conditions within a reservoir.
• Make sure you alter these values to suit your particular reservoir. Try to make
the region that is defined by your values as small as possible. For example if
you think that the maximum temperature in your reservoir is going to be 300F
then do not leave the default value in the panel (400F).
• If the region you define is too big then Crookston’s formula may have trouble
getting good fits to the EoS based K-values. This is because K-values are really
a function of the fluid composition as well as the pressure and temperature.
See "K-Values" on page 423 for an explanation of why this is so.
3 The number of flashes variable is set by default to 20. PVTi randomly throws this
number of points into your defined region and flashes are performed at each (P,T)
coordinate in order to obtain K-values. These are the observations. You are
probably fine using 20 as this value. If your region is particularly big (which is not
recommended) then you may want to increase this. By the same token, decreasing
this value will speed up the calculations significantly, but the fits may be less
accurate.
4 The A and D coefficients are turned on by default. You can turn them off if you
want to but experience shows that you will almost always need them to get the best
fits to the observations. In theory, turning all of the coefficients on should give the
best fits. In practice, although you will do this most of the time, you will find:
• this slows the application down as it increases the parameter space to search.
If you want to get a feel for how the module works start with A, B and D turned
on.
• the algorithm finds the nearest local minimum of the objective function. This
may not be the global minimum. PVTi uses an intelligent guess and the idea is
that it starts in the parameter space reasonably close to the global minimum so
that this is the nearest local minimum. When all 5 coefficients are switched on
this initial guess technique is not always guaranteed to give as good a guess as
when less coefficients are selected.

5 When you view the fits, check to see how good they are when you use a constant
value of temperature that is halfway between the maximum and the minimum
one. In theory the pressure part of Crookston’s equation should show the best fit
here. If the functional form of the line plots looks wrong for even a single
component then you probably need another pressure coefficient in order to match
all the components for this particular fluid. Check in the same way for the
temperature part of Crookston’s equation and turn E on as well if the D coefficient
is having trouble fitting the observations
6 In terms of the rms values you should expect in the fits, this depends on how large
a region you define and how many components are in your fluid. If the region is,
say 1000 psia high and 200K wide, then you should expect fits on the observations
of 2-3% on average.
For more technical information on the ECLIPSE Thermal export facility see "ECLIPSE
Thermal Export Module" on page 423.

Analysis techniques
Introduction
• "Recommended PVT analysis for oil reservoirs" on page 393.
• "Recommended PVT analysis for gas condensate reservoirs" on page 398.
• "Consistency tests and correlations" on page 403.

Analysis techniques
Recommended PVT analysis for oil reservoirs
Oil reservoirs
Oil reservoirs have been divided into three classes in the following discussion:
Under-saturated reservoirs
Refers to those reservoirs that are under-saturated not only at the time of their
discovery but also throughout their development.
Note This class does not include reservoirs whose pressures fall below the bubble
point at some time during their development, or reservoirs that are subject to
gas injection.
Under-saturated reservoirs can be represented by two-phase (oil and water)

simulation models, usually using the black oil formulation. For further information see
"Under-saturated reservoirs" on page 395.
Saturated black oil reservoirs

Refers to those reservoirs whose pressures fall below the bubble point somewhere in
the oil column during development, but that can be modeled sufficiently accurately
using the black oil formulation.
Note All reservoirs with initial gas-caps are regarded as saturated (but may not be
adequately represented by the black oil formulation).
For further information see "Saturated blackoil reservoirs" on page 396.
Volatile oil reservoirs

Refers to those reservoirs in which phase behavior effects are important. These
generally require extensive use of a PVT program and compositional modeling for
accurate representation.
Note Volatile oil usually refers to a reservoir fluid whose critical temperature is only
slightly above the reservoir temperature under initial conditions.
For further information see "Volatile oil reservoirs" on page 397.

Black oil or compositional simulation
You can generate both black oil tables (for black oil simulation) and equation of state
data (for compositional simulation) using the PVTi package. In some cases you can
generate black oil tables from laboratory results alone, rather than using the full
equation of state model. See "COMB - Compositional Material Balance" on page 123
and "COMB section keywords" on page 170.
Oil reservoirs
Almost all under-saturated reservoirs, and most saturated oil reservoirs where the
reservoir fluid has a critical temperature far above the reservoir temperature, do not
normally require the complexity of compositional simulation, and a black oil
formulation is normally of sufficient accuracy.
Note Such a formulation does not allow the stock tank oil density or gas density to
vary with time.
You may wish to consider using compositional simulation if miscible or semi-miscible

processes are involved.

For volatile oil reservoirs it may be possible to obtain some sort of approximation to
reservoir fluid behavior during depletion using a black oil formulation, providing that
this has been modified to allow the gas phase to contain some vaporized oil (variable
R v ).
In general, however, a compositional simulation is required for accurate treatment.

This applies even for depletion of the reservoir.
A compositional approach is essential for reservoirs where gas injection is planned.
Compositional grading with depth

Many oil reservoirs exhibit some kind of grading of fluid properties with depth.
If you use the black oil formulation, then you can often represent the effects of this
grading by a variation of solution gas-oil ratio with depth.
However, if the stock tank oil API varies with depth, and it is necessary to model the
variation of API with time, then you should use a variable oil API option in the black
oil runs.
Where part of the fluid column is near critical conditions you would use an initial
compositional gradient with depth in a compositional simulation.
Some reservoirs may grade from a gas condensate to a volatile oil without an initial
phase transition or contact of any kind. Others may have gas-liquid or even liquid-
liquid interfaces.
You can simulate this process with the COMPG experiment in PVTi. One of the
quantities output to the PVP file is the API, which you then use as the basis for
constructing the API versus depth variation. For further information see "Defining
Experiments" on page 128.

Under-saturated reservoirs
For under-saturated reservoirs, it is not often that an equation of state representation is
of particular value, although you may find the PVTi program useful for separator
calculations. Although it is possible to derive black oil tables from equation of state
models, for this application it is better to derive them from laboratory data for the
following reasons:
Laboratory analyses generally have sufficient information to derive the PVT tables
directly.
It is often difficult to characterize reservoir oils accurately. This leads to difficulty in
establishing an equation of state model that satisfactorily matches the measured data.
To establish a black oil PVT model of a single sample from an undersaturated reservoir
the following procedure is recommended:
1 From the laboratory report take the saturation pressure (bubble point) of the
sample, the reservoir density at that pressure, and the compressibility (as a
function of pressure) above the saturation pressure.
2 If the field is under production there is normally data available on the producing
GOR for the field separator conditions, the gas gravity ( γ g , air = 1), and the stock
tank oil gravity (API).
3 Ensure that the GOR and gas gravity include all the separator and stock tank gases.
If they do not then you can often make reasonable estimates. You can then combine
this data with the reported oil density to give the formation volume factor ( B o ) at
the saturation pressure p b as follows (in field units):
1 141.5 γ g ⋅ GOR ⋅ 0.00122

B o ( p b ) = --------- ---------------------------------- + ---------------------------------------------
- [EQ 8.212]
ρ res ( 131.5 + API ) 5.6146
where ρ res is the reservoir density in gm/cc, and the GOR in scf/rb.
4 You can find the oil formation volume factor at pressures higher than the
saturation pressure using the compressibility factor in the PVT report.
5 If the field has only been tested and there is no data on producing GOR and gas
gravity then often a separator test for appropriate separator conditions are
reported. This supplies a formation volume factor for bubble point oil at these
conditions of separation. You can use the reported compressibilities to derive
formation volume factors at higher pressures.
If a separation test is reported and the field is under production, then you should
rationalize the formation volume factors derived from point 2 with the separator
test.
6 If no separation test was performed by the laboratory and there is no production
data, then you can either use a suitable set of correlations, or, if the sample is a re-
combination, it may be possible to use the re-combination data to derive data at
separation conditions using PVTi, providing it is possible to characterize the fluid.

Note It is important to avoid using what is often referred to as the formation volume
factor (it is actually a relative volume) during a differential liberation
experiment. The oil remaining at the end of this experiment has a completely
different composition from stock tank oil and this formation volume factor is
usually higher (by as much as 10%) than the correct value.
For the simulation of undersaturated reservoirs this differential liberation
experiment is usually superfluous, although laboratories often perform it
anyway.
If there are a number of samples you should follow the above procedure for each one.
7 Then determine if there evidence of an API gradient by plotting stock tank oil
density against depth.
8 If there is no such evidence take the most appropriate sample (or average of
samples). This becomes the PVT representation of the reservoir oil.
If the reservoir has an API gradient the following additional stages are recommended:
9 Draw a representative curve through the API versus depth plot. Use this to
represent the model API versus depth relationship of the reservoir oil.
10 There are a number of ways of allocating PVT properties as functions of API
gravity. By suitable plots, such as saturation pressure against depth, formation
volume factor for a common pressure against depth, and formation volume factor
against pressure, you can find a method appropriate to that particular reservoir.
You may find it necessary to use correlations to extrapolate to depths above and
below the range of sample data.
Saturated blackoil reservoirs

Most of the stages for "Under-saturated reservoirs" on page 395 are appropriate for
these reservoirs.
In addition you must generate a table of solution gas-oil ratio, oil and gas formation
volume factor, and oil and gas viscosities below the bubble point.
Normally, you can derive this data from the separation test and differential liberation
experiments performed in the laboratory.
Note Again it must be stressed that you should not use differential liberation data
alone. What is often referred to as the formation volume factor during the
liberation makes no reference to separator conditions.
To be completely accurate the laboratory should perform a separation test on the fluid
resulting from each pressure stage of the differential liberation experiment.
In practice this is rarely done, and the volumetric data is determined from the single
separation test (at the bubble point pressure), and the relative volumes and gas
volumes of the differential liberation experiment.

The correct oil formation volume factor is this relative volume, divided by the relative
volume of the bubble point fluid (during the same differential liberation), multiplied
by the correct oil formation volume factor of the bubble point fluid for the separator
conditions operating. This derivation, of course, uses the black oil approximation that
the stock tank oil density is unchanging.
Solution gas-oil ratio is derived in a similar way, the value at the bubble point pressure
being exactly that derived from the separator test ( R s ( p b ) ).
Simple material balance of gas and oil shows that the correct gas-oil ratio ( R s ) at a
pressure below the bubble point is given by:
R s = R s ( p b ) – ( R s' ( p b ) – R s' )B o' ( p b ) ⁄ B o ( p b ) [EQ 8.213]
where R s' and B o' refer respectively to the reported solution gas-oil ratio and
formation volume factor from the differential liberation experiment.
As well as API plots against depth for different samples, plots of producing gas-oil
ratio and formation volume factor against depth (for a consistent pressure) may
indicate a solution gas-oil ratio for a such a reservoir.

A differential liberation experiment, or a constant volume depletion experiment, or
both, may have been performed on a volatile oil sample. In practice, during depletion,
some intermediate process occurs in the reservoir because gas migrates away from the
oil, although it might be argued that enough gas will be trapped near the oil for a
constant volume depletion experiment to be most representative. A swelling test may
also have been performed if gas re-injection is planned.
You should subject the results of any constant volume depletion experiment to the
same analysis as for a gas condensate fluid sample, as described in "Recommended
PVT analysis for gas condensate reservoirs" on page 398.
The objective of the PVT analysis is to generate an equation of state model of the
reservoir, which you then use for compositional simulation. If it is felt that the
depletion experiment is a good enough representation then you can generate a black
oil table from the equation of state model.

Recommended PVT analysis for gas condensate
reservoirs
Introduction
• "Blackoil or compositional simulation" on page 398.
• "Data analysis" on page 399.
• "Compositional grading with depth" on page 399.
• "Data analysis" on page 399.
• "Equation of state model" on page 400.
Blackoil or compositional simulation

For most gas condensate reservoirs the liquid that condenses in the pore spaces during
depletion does not reach a high enough saturation to become mobile. Gas and oil do
not move with respect to each other, and they therefore stay in equilibrium. It is
possible to model this type of reservoir behavior quite accurately by the constant
volume depletion (CVD) experiment performed in the laboratory.
For further information on defining experiments in PVTi see "Defining Experiments"
on page 128.
For a straightforward gas condensate, a black oil formulation, modified to include a
variable R v vapor oil-gas ratio, can therefore do a reasonable job of modeling simple
depletion.
Although the black oil formulation nominally constrains stock tank oil and gas to have
unvarying compositions and densities, it is possible to reproduce the compositional
data available for the original constant volume depletion experiment for surface
facility calculations by post-processing black oil output.
However, if the gas condensate becomes very rich, perhaps grading to a volatile oil,
then it requires a fully compositional simulation to model it.
This is also the case when you wish to consider almost any other development scenario
than depletion, except, perhaps, pressure maintenance by water injection.
Dry gas cycling, in particular requires fully compositional modeling for accurate
predictions of the phase effects when dry gas contacts condensed reservoir oil. The
extended black oil formulation allows dry gas to pick up oil until the gas becomes
saturated, an optimistic approximation to the actual reservoir behavior.
When you generate black oil tables from a CVD experiment by PVTi, the program
includes dry gas properties in the table, allowing gas cycling to be modeled.
Note The results of such modeling are to over-estimate the condensate recovery
from the field by a few percent for full and partial pressure maintenance. Low
pressure recycling will probably be very poorly modeled.

Recommended PVT analysis for gas condensate reservoirs
Compositional grading with depth
Where part of the fluid column is near critical conditions an initial compositional
gradient with depth would be used in a compositional simulation. Reservoirs may
grade from a gas condensate to a volatile oil without a phase transition or contact of
any kind. More often there is an oil-rim, which can exist even without critical
conditions being approached. For further information on defining experiments in PVTi
see "Defining Experiments" on page 128.
Provided reservoir fluids are not close to critical conditions, a reasonable
approximation to recovery can often be made using a black oil formulation, by
modeling the dew point variation and ignoring the condensate-gas ratio. The black oil
formulation cannot normally reproduce, simultaneously, the variation in dew point
pressure and in condensate-gas ratio.
Data analysis
A common failing when analyzing gas condensate reservoirs is to attempt to establish
an equation of state representation without thorough analysis of the data on which it
is to be based.
A thorough comparison of samples, analysis of inter- and intra- sample consistency,
and clarification of the aims of establishing an equation of state model, allows you to
determine which data is worth trying to match, what components to use, and how best
to establish a matched model.
You should closely scrutinize all the available samples. Some of the questions that
should be answered are as follows:
• Are the samples properly constituted (examine recombination data and compare
with correlations)?
• Have experiments been performed consistently and accurately (examine
compositional material balances)?
• Are other experimental data such as Z -factors, implied K -values and liquid
densities reasonable (correlations and tests of consistency)?
• What are the characteristics of the heavy fractions (fingerprint plots)?
During this analysis it may seem reasonable to establish alternative (modified) data by
adjusting sample data to achieve material balance consistency and a reasonable fluid
description. This can be a useful exercise but a good appreciation of the main sources
of inaccuracy is required for it to be successful.
If the sample is a recombination, the report usually contains recombination data that
can be used for calibrating correlations, to obtain an analysis of separation using the
COMB section of PVTi. For further information see "COMB - Compositional Material
Balance" on page 123 and "COMB section keywords" on page 170.
It is possible to obtain gas and condensate recovery estimates for different
abandonment pressures by combining this with constant composition data above the
dew point, and CVD data below the dew point.
After you have analyzed each sample individually you should examine all the samples
together by, for example, plotting dew point pressures and condensate-gas ratios
against sampling depth, comparing compositions, fingerprint plots etc.

Before rejecting any particular sample that appears to be anomalous, you should
investigate the source of anomalies. Such an investigation may be useful in making
recommendations concerning sampling techniques or laboratory analysis.
Equation of state model

When a good overall view of the reservoir field has been gained, it is then possible to
establish PVT models. If depletion is planned it may not be necessary to establish an
equation of state model. If the basic (or modified) laboratory data is of sufficient quality
you can use to establish black oil tables. Normally, however, you establish equation of
state models.
For further information on fitting data to an EoS in PVTi see "Equation of State" on page
111.
Different engineers have their own methods of establishing an equation of state model,
but a suggested approach is as follows:
1 Establish a model based upon as many components as the data defines, using
correlations where necessary. Split the last component (the plus fraction) into, say,
three further fractions, using one of the splitting options. The modified Whitson
(Semi-Continuous-Thermodynamics (SCT)) is recommended. For further
information see "Multi-feed Split (also called semi-continuous thermodynamic
(SCT) split)" on page 117.
2 Previously, it was recommended that the binary interaction coefficient between
Methane and the plus fraction be varied to match the dew point pressure. This
practice is no-longer recommended. Adjusting binaries to match saturation
pressure at just one temperature often massively distorts the phase envelope at
other (generally lower) temperatures, see [Ref. 30].
Instead, it is now recommended to use one of the following approaches:
• If using Cheuh-Prausnitz BIC’s, try the (pre-multiplying) A -coefficient. This
adjusts ALL hydrocarbon-hydrocarbon binaries by the same amount,
maintaining symmetry and monotonicity. For further information see
"Viscosity correlations" on page 112.
• Having performed an SCT-split, regress on the plus fraction mole weight,
PVTi’s internal correlations then convert this modified mole weight into
critical properties, etc., which are monotonic and consistent.
• If the above are unavailable, critical temperature of plus fraction usually
requires changing less than any other individual EoS parameter.
3 Compare this first model with the data to be matched. This usually comprises the
volumetric data associated with constant volume and constant composition
experiments, separator data and swelling test data. As well as the volumetric data,
it is important to compare compositional data, although the possible inaccuracies
in reported compositions should be considered here.
4 Attempt to establish a model that is a good match with only a minimal
modification of the data. Such modifications may be made manually or semi-
automatically but you should obtain a good understanding of the effects of each
modification.
The emphasis is on the modification (preferably consistently) of individual
component descriptions, especially those of the heavy components which are
poorly defined. The splitting of the plus fraction may be a suitable modification to

make.
Other properties to consider are the critical temperature and pressure of the plus
fraction along with its acentric factor. Additionally, because of the generally
uncertain Paraffinic-Napthanic-Aromatic (PNA) of the Single Carbon Number
(SCN) groups, that is hexanes, heptanes, etc., these are a better option for a more
limited change.
If such a minimal set of modifications does not give a satisfactory match then a full
regression procedure may be necessary as described for pseudoization below. Ideally
this full regression will not be necessary before pseudoization.
1 Generate a black oil table from this many-component model from a simulated
CVD experiment. For further information on defining experiments in PVTi see
"Defining Experiments" on page 128.
2 Decide upon an appropriate set of pseudo-components, bearing in mind the
objectives of the study. Use the grouping option to generate the pseudo-
component model from the many-component model. In many cases about six
pseudo-components is appropriate.
3 Compare the predictions of the pseudo-component model with the laboratory
data.
4 Use the automatic regression procedure to modify the representation. The
emphasis now is on obtaining a good match to the data and the results of the
many-component model. Fairly coarse matching parameters may be required.
Modify parameters for which a small change has a large effect. (The matrix of
sensitivities produced by the regression algorithm may be useful here.) You should
try several sets of regression parameters and you should compare the predictions
over the whole area of interest (including mixing of dry and wet gases) with the
predictions of the many-component model and with the data (where available).
When pseudoizing, the ideal grouping results in minimal predictive change from
the original system. This is not always possible and some small changes in the
group properties may be necessary.
5 If there are problems matching reservoir data and separator data simultaneously
then it is possible to divorce the two in ECLIPSE for compositional simulation,
using different Ω values and binary interaction coefficients for the separators, or
using K -values for the separators.
It is also possible to account for temperature variation using the temperature-
dependent volume shift. By matching separator volumetric properties first, that is
at or near surface temperature where the volume shifts are actually defined, you
can then attempt to match to reservoir volumetric properties, at the elevated
reservoir temperature, by using the thermal expansion coefficient that is available
as a regression variable.
One advantage of the above procedure is that it addresses directly one of the main
problem areas in this type of work: the characterization of the heavy fractions.
These components have an effect on fluid behavior that far outweighs their sometimes
small mole fractions. You always rationalize any splitting performed with any true
boiling point (TBP) distillation data available, fingerprint plots, correlations etc.
A second advantage of the recommended procedure is that it allows (through the
many-component model) a good understanding of the character of the fluid.
A third advantage is that if it is possible to obtain a good match to the laboratory data
with minimal modification of the pure many-component equation of state model then
it is more likely that accurate predictions are obtained away from the measured data.

Note A cubic equation of state can only supply approximations to the behavior of
fluids. This must be borne in mind when matching laboratory data, but
especially when extrapolating outside the area of data control. By the same
token, laboratory analyses of samples only normally supply an approximation
to the sample behavior and the sample itself is only an approximation to the
reservoir fluid.

Consistency tests and correlations
Introduction
During initial data analysis a number of tests are available in the COMB (COmpositional
Material Balance) section of the program for examining the consistency and quality of
a particular laboratory sample.
For further information see "COMB - Compositional Material Balance" on page 123 and
"COMB section keywords" on page 170.
This section contains information on the following:
• "Compositional material balances during the CVD experiment" on page 403.
• "Fluid density correlations" on page 403.
• "Viscosity calculations" on page 404.
• "Separator calculations" on page 404.
• "K-value checks" on page 404.
Compositional material balances during the CVD

experiment
From the volumetric and gas compositions reported during the various pressure stages
of the CVD experiment it is possible to make deductions about the composition and
density of the remaining liquid at each stage. For further information on defining
experiments in PVTi see "Defining Experiments" on page 128.
If the moles of liquid remaining after the last stage is reported (which, unfortunately,
is not always the case) it is possible to combine this with the fractions recovered at each
stage to compare the initial and recovered compositions.
This overall comparison can be revealing in terms of laboratory accuracy and the
desirability of making modifications to the basic data before making a match.
Fluid density correlations

It is possible to compare the liquid density calculated from compositional material
balance with calculated values using the Alani-Kennedy correlation, [Ref. 22].
Several correlations are available for the gas phase.
It is possible to compare the pseudo-critical temperature and pressures (the internal
correlation uses the Wichert-Aziz correction for sour gases). You can compare these
pseudo-critical properties with those calculated using a correlation based on the
hydrocarbon molecular weight.
It is possible to compare a representation of the Standing-Katz Z -factor correlation by
Dranchuk et al., [Ref. 21], using the Benedict-Webb-Rubin equation of state with
laboratory measured Z -factors.

K-value checks
The Hoffmann-Crump-Hocott (HCH) technique consists of plotting the logarithm of
the product of the K -value and the pressure against a characterization factor for each
component.
HCH found that at any given pressure this data plotted close to a straight line, at least
for the pure components. It is possible to check K -values derived from the (calculated)
liquid and (reported) gas compositions during the various stages of the CVD
experiment. For further information on defining experiments in PVTi see "Report" on
page 124.
This technique is also useful for checking the recombination of samples. It forms the
basis of Standing’s method for performing separator calculations, discussed later.
Another good check of the K -values from a CVD experiment is to plot them against
the logarithm of pressure for each component. This allows an estimation of the
apparent convergence pressure, which then allows cross-checking against NGAA
tables.
Viscosity calculations
Laboratories generally do not have the equipment for measuring gas viscosities and
usually calculate viscosities based on the Carr, Kobaysahi and Burrows correlation.
Other correlations exist, such as the "Lohrenz, Bray and Clark" on page 345 method and
the "Pedersen et al." on page 347 corresponding states technique. These different
correlations often yield markedly different predicted viscosities.
Ideally, a consistent correlation should be used throughout the reservoir engineering
analysis, from the well test analysis used to derive permeabilities to the reservoir
simulation. ECLIPSE accepts a set of critical volumes (entered with VCRITVIS or
ZCRITVIS), which are only used for viscosity calculations, in order that significant
changes in V c values used to match viscosity correlations do not effect the remainder
of the simulation.
Separator calculations
Standing evolved a method for calculating K -values for separators, based on the HCH
method. For each component, a b -factor is calculated at the separator temperature.
HCH suggested that this b -factor is based on pure component normal boiling points
and critical temperatures.
Standing suggested a slightly different set but this probably only applied to the
particular oil he was considering. b -factors obtained by the HCH method may be used
in PVTi to generate separator K -values.
An alternative is to use those derived from recombination data, though these should
first be checked against those obtained using the HCH method.

From the K -values derived in this way, PVTi can calculate the condensate and oil
recoveries for a given separator system. Separator liquid densities are based on the
Alani-Kennedy correlation (see "Alani-Kennedy liquid densities" on page 324)and
stock tank densities are calculated using Amagat's law (addition of specific volumes).
Another method of calculating recovery factors is to assume stabilized liquid consists
of, say, the Pentanes plus fraction, whilst the stabilized gas consists of the other, lighter
components.
With these assumptions, recoveries down to the last pressure in the CVD experiment
are commonly reported in laboratory analyses in units of gallons of condensate per
mscf of wet gas. Laboratories calculate these using Amagat’s law. You can perform
these calculations in PVTi using input specific gravities and molecular weights.
From input CCE, CVD data and the separation system, PVTi can calculate gas and
condensate recovery factors from any supplied initial reservoir pressure to any
supplied abandonment pressure, for a depletion scenario.

Introduction
Fluid properties estimation is used in situations where a full equation of state matching
across numerous samples is not available.
Typically, there is only one sample and limited information about the fluid behavior.
From this, fluid properties estimation simulates typical experiments and provide
analysis of the fluid ahead of full laboratory experiments.
Minimum information
In fluid property estimation the following data provide sufficient information to
construct a complete set of example experiments:
• Weight percentage of each component (or mole fractions)
• The mole weight of the plus fraction
• A salutatorian pressure (bubble or dew point)
• The maximum pressure to use when constructing a depletion experiment (the
default is to use the saturation pressure.
Fitting to saturation pressure

The plus fraction mole weight is used to characterize the critical properties of the plus-
fraction. In general, the weight fraction of the plus-fraction is well known, but the
appropriate mole weight to use in characterizing the critical properties is not.
By varying the weight of the plus-fraction, whilst maintaining the weight fraction
constant, PVTi is able to determine the most appropriate mole weight to use in
characterizing the critical properties. This gives a good fit to the entered saturation
pressure and creates a fluid model that can be used for fluid properties estimation.
Generation of pressure depletion experiments

In generating pressure depletion experiments, PVTi begins by creating the Constant
Composition Expansion (CCE). There are up to ten pressure steps from the maximum
entered pressure down to the saturation pressure, and up to ten pressure steps from the
saturation pressure down to standard pressure. If no maximum pressure was
provided, the CCE starts at the saturation pressure.
If the saturation pressure was a bubble point, a Differential Liberation (DL) experiment
is created, otherwise a Constant Volume Depletion (CVD) experiment is created. The
DL or CVD has the same pressure steps as the CCE previously created.

Optimized separator
An optimized separator calculation can be performed. The method for this is discussed
in "Optimized separators" on page 361.
Technical Information
The default settings for the PVTi fluid properties estimation are in Table 8.13.
Table 8.13 PVTi defaults for Fluid Property Estimation
Property Value
Equation of state Peng-Robinson 3-parameter (corrected)
Library for component properties Katz-Firoozabadi
Correlations for plus-fraction properties Kesler-Lee
Viscosities Lohrenz-Bray-Clark
The equation-of-state and the viscosity model can be changed using PVTi: Edit | Fluid
Model | Equation of State...
The component library in use can be selected in PVTi: Utilities | Program | Options...
The correlation used for the plus-fraction can be changed in PVTi: Edit | Fluid Model |
Components...

Regression in PVT analysis
Introduction
The goal of PVT analysis is to provide a tuned Equation of State that can model the
reservoir fluid in simulations. The selection of regression parameters is crucial in
determining the quality of the tuned fluid model.
This chapter discusses the selection of regression variables in PVT analysis and
consolidates traditional variable selection, based on physical principles ("Physical
selection of regression parameters" on page 408), with variable selection based on
mathematical analysis of the problem ("Mathematical analysis of the regression
problem" on page 410). "A consistent methodology that can be applied automatically"
on page 411, describes how information in the preceding two sections are used by the
PVTi Quick Fit option.
Fundamentally, the fitting process twists
the standard cubic-equation of state to fit the supplied data. Across the temperature
and pressure ranges of the observations the fitted model may be useful, at very
different temperatures and pressures the model may not be representative of the fluid
behavior at all.
Note Always use data at reservoir conditions, such as depletion experiments, and at
surface conditions, such as separator tests, when fitting a fluid model that is to
be used in reservoir simulation.
In "Weighting observations for regression" on page 412, there is a discussion on

weighting the observation data to get the desired fit.
This chapter does not contain a mathematical description of the Levenberg-Marquardt
algorithm, that can be found in "Weighting observations for regression" on page 412.
Physical selection of regression parameters

The traditional, chemistry-centered approach to equation-of-state fitting relies on first
determining which parameters are least well defined. Parameters that are not known
accurately are candidates for fitting.
Critical properties
For pure library components, particularly the non-hydrocarbons and lighter
hydrocarbons, the acentric factor (ω) and critical temperature (Tc) and pressure (Pc) are
well known. It is generally not justifiable to tune these properties.
Hint It is very unlikely that the Pc, Tc and ω of H2O, H2S, N2, CO2, CO, C1, C2, C3,
IC4, NC4, IC5, NC5, and C6 differ from the library values. So do not use them
in regression.

The critical properties of the heavier components are less well determined. The plus-
fraction is a mixture of many different hydrocarbons and usually the properties are
determined by correlations based on mole weight and specific gravity of the plus
fraction as a whole. Consequently the critical properties are only as good as the
characterization method that generated them.
Hint The Pc, Tc, and ω of the plus fraction usually come from characterization. This
means that they are not well determined, making them ideal candidates for
tuning.
Pedersen et al. have suggested that the molecular weights of the characterized
components might be a good choice of regression variable as this varies Pc, Tc and ω
consistently, in line with the chosen characterization method, see [Ref. 60].

The binary interaction coefficients are introduced into the cubic equation-of-state to
account for the effect of polar forces in the interactions between components. This
means that for non-polar interactions, such as those between hydrocarbons, the binary
interactions are close to zero.
There are lots of binary interaction coefficients, generally there is insufficient PVT data
to justify tuning all of them. Over-fitting, fitting the data more closely than is justified,
can lead to a fluid model that is not physically valid outside the range of the data being
fitted. In the reservoir simulator you model many compositions outside the observed
PVT data-set. Over tuning, particularly using the binary interactions, can cause serious
convergence problems.
Caution Careless tuning of binary interaction coefficients often leads to a fluid

model that has convergence problems in compositional simulations.
The basic cubic equation-of-state uses the acentric factor to consider slight deviations
in molecule shape; the assumption is that all molecules are close to being spherical. The
binary interaction coefficient between light components and the heaviest hydrocarbon
can be used to compensate for the non-sphericity of the heavy hydrocarbon molecules.
ΩA and ΩB coefficients
The default values of ΩA and ΩB are based on the following assumptions:
1 The critical isotherm of a component has a slope of zero and an inflection point at
the critical point. In layman’s terms this means that the component is pure.
2 That the determination carried out, for the limited number of pure components
available, can be extrapolated to all heavier and lighter components.
3 The values were valid for pure-component density and vapor pressure below
critical temperature.
In developing his equation of state, Tareq Ahmed explained why these basic
assumptions are violated in many fluid models, see [Ref. 62]:
1 Pseudo-components and the plus fraction are not pure components, so assumption
1 is violated by these.

2 The heavier hydrocarbons and light components may not have the same ΩA and
ΩB as the pure components tested. Specifically, assumption 2 might not be true.
3 Light components such as methane and nitrogen are well above their critical points
at temperatures typically encountered in PVT studies. Assumption 3 does not
apply to these components.
Hint The assumptions upon which the default ΩA and ΩB constants are based do
not hold for light components, pseudo-components, and the plus fraction. This
makes them candidates for regression.
Lohrenz-Bray-Clark viscosity coefficients

The Lohrenz-Bray-Clark viscosity correlation is a fourth order polynomial in density.
Consequently it is very sensitive to the density of the fluid. In PVTi there is the option
to regress on the critical volume or Z-factor used in calculating each component’s
contribution to the total viscosity.
Hint Varying the critical volume and Z-factor are equivalent in this case, so it is
futile to vary both at once.
This critical volume or Z-factor is only used to calculate the viscosity, so remember that
in a regression only the viscosity observations themselves are used to tune these
parameters.
Caution As only a few observations apply to the tuning of the critical volumes for
viscosity, make sure that there are no more than one or two regression
variables of this type in a regression run. It is much better to put all critical
volumes for viscosity into one regression variable than to vary all of them
at once.
There is another option in PVTi that allows the constants used in the fourth order
polynomial in density to vary.
Caution This option is available so that the original Lohrenz-Bray-Clark analysis

can be repeated with new data-sets. It should only be used if you have a
large number of samples and lots of viscosity observations. Fitting too
many parameters to too few observations leads to convergence problems
for the fluid in a compositional simulator.
Mathematical analysis of the regression problem

In this section the analysis of the regression problem is based upon the need to create
a solvable mathematical system of equations. The physical reasons for changing
parameters were laid out in the previous sections, this defined a set of candidates for
regression. The mathematical analysis indicates which candidates should be regressed
together in a single regression run.

Understanding the Levenberg-Marquardt algorithm
The Levenberg-Marquardt algorithm, like many non-linear regression algorithms,
starts from the assumption that the non-linear problem can be solved as a set of small
linear steps. This, in effect, means that the problem must not be too non-linear. The
regression can be helped by designing the problem, that is choosing regression
variables, such that the problem is not too non-linear. What is meant by too non-linear?
It is more straightforward to start from the linear problem and bend the rules a little.
The characteristics of the linear problem are:
1 No regression parameters are numerically redundant
2 All regression parameters are independent
High sensitivities
The first rule means that interest should focus on the most sensitive parameters, those
for which a small change has a large effect on the fit.
The Hessian tab of the Sensitivity Analysis panel gives an indication of the sensitivity
of the different parameters. The values along the leading diagonal of the Hessian
matrix are most useful, a high value means high sensitivity.
Hint By removing insensitive regression variables from the set, the problem
becomes less non-linear and thus more solvable.
Kai Liu has reported success with an automatic regression technique based solely on
selecting the eight most sensitive regression parameters, see [Ref. 62].
Correlations between parameters

The second rule means that there should be no very strong correlations between
parameters. A strong correlation means that the effect of a change in one parameter is
the same as, or opposite to, the effect of changing another.
The Correlation Sensitivity Analysis panel shows the correlations between parameters.
A value close to 1 indicates a strong correlation, which means changing one parameter
has the same effect as changing the other. A value close to -1 indicates a strong anti-
correlation, which means changing one parameter has the opposite effect to changing
the other.
Hint Amalgamating strongly correlated regression variables of the same type (for
example Tc) in a single regression variable or removing one of the two strongly
correlated variables makes the problem less non-linear and thus more
solvable.
A consistent methodology that can be applied

automatically
This section describes how the rules laid out in "Physical selection of regression
parameters" on page 408 and "Mathematical analysis of the regression problem" on
page 410 can be applied to a automated regression system.

Note By their nature automatic regression schemes do not make judgement calls
that an experienced engineer would make.
The choice of regression variables

From the rule-of-thumb guidelines given in "Physical selection of regression
parameters" on page 408, the following set of properties was determined as begin
available candidates for regression:
1 Tc, Pc and ω of any non-library component.
2 Tc, Pc and ω of any component with mole weight of C7 or heavier. (As these are
effectively mixtures of different molecule types and so may differ from library
values).
3 ΩA and ΩB of any component with mole weight of C7 or heavier. Again because
these are mixtures.
4 No binary interaction coefficients because of the risk of over-fitting.
5 No viscosity-specific parameters, again because of the risk of over-fitting.
This gives a large number of regression variables, many of which are very closely
interrelated. The next step is to apply a mathematical analysis onto this set.
Reducing the set of regression variables

From "Mathematical analysis of the regression problem" on page 410 it is clear the
problem can be made more solvable by selecting those parameters with high
sensitivities and low inter-correlations.
From the discussion, the following rule was created:
• Regress using the set of variables with the highest total sensitivity, for which no
correlation between parameters is greater than 0.9 or less than -0.9; and for which
the lowest sensitivity is at least 1% of the highest sensitivity.
The total sensitivity of the set of variables was taken to be the sum along the leading
diagonal of the Hessian. The sensitivity of a regression variable was taken to be the
value on the leading diagonal of the Hessian relating to that variable.
Note This method takes no advantage of amalgamating regression variables of the

same type into a single variable.
Weighting observations for regression

The least-squares fit to the observation data is not necessarily the goal of equation-of-
state fitting. To be precise, the goal is to create a fluid model that behaves like the
reservoir fluid. There are aspects of the reservoir fluid that the model must capture
accurately.
PVTi allows the weighting of regression variables so that additional importance can be
attached to some measurements over others.

Matching saturation pressure
Usually, it is important to match the fluid’s saturation point within the depletion
experiment. This value is closely related to the gas-oil contact in the reservoir and can
strongly affect the initial fluids in place calculation.
Hint Use the saturation pressure measurement from the depletion experiment to
create a dew-point or bubble-point experiment. Weight the observation of
saturation pressure very high (1000+).
Matching surface densities or GOR

The separator densities or GOR strongly determine the total oil and gas produced. Also
it is worth remembering that the depletion experiment often has 20 or more stages,
whereas the surface separator is unlikely to have more than 2 or 3 stages. This means
there are usually lots more observations for the reservoir depletion process than for the
produced oil and gas.
Hint Weight the separator densities or GOR highly (500+) to guarantee that the right
amount of oil and gas are produced.
Guiding the regression

Sometimes a fluid does not seem to fit the observations. In this situation, the regression
weights can be used to guide the regression algorithm.
For example, an excellent fit may have been achieved for the bubble point using the
method described in "Matching saturation pressure" on page 413, but there may be
difficulty in fitting the liquid saturation below the bubble point.
PVTi allows the weighting of individual observations for this purpose.
In this situation, by weighting the observations for which the fit is worst by the highest
amount, the regression is forced to take more account of the data as a whole.
Hint Weight the worst-fit observation the most and successive observations less and
less through to the best fit observations that retain their original weighting of
1. This guides the regression to take more account of the worst parts of the fit.
Caution Be aware of the errors in the observed data and do not be tempted to over-
fit. Models that are over-fitted cause convergence problems in
compositional simulations.

SmartOpt - Systematic MAtrix Reduction
Technique for Optimization
This algorithm performs a rapid and exhaustive search of the sensitivity matrix
described in "Reducing the set of regression variables" on page 412. As the non-linear
regression is, effectively, a succession of linear steps, the problem must be designed
such that it is only weakly non-linear.
The number of combinations of regression variables that could be used for regression
is 2N where N is the number of regression variables. For even 20 regression variables
this would be 1,048,576 combinations, for 30 variables the number of combinations is
more than a billion. To calculate the correlations between parameters, a matrix
inversion is required, clearly several million matrix inversions is impractical. By re-
designing the problem, however, a fast and efficient exhaustive search can be made to
find the most sensitive regression matrix representing a weakly non-linear problem.
A starting point - the most sensitive single variable

If the run is made with just one regression variable, there are no other variables so
correlation is not an issue. The starting point then is to find the single most sensitive
variable. This gives us an initial estimate of the best sensitivity - no combinations with
lower sensitivity than this need be tried.
Ordering for speed

The Hessian is reformed with the most sensitive variable in the first row, down to the
least sensitive variable in the final row. As is explained in the next section, this gives an
important speed increase.
Searching all combinations

The speed in this method relies reducing the time spent analyzing combinations that
do not turn out to be the best one.
The first test to be applied, then, is the sensitivity difference - no parameter should have
less than 1% the sensitivity of the most sensitive parameter. As the combination is
constructed, it is tested against this criterion. If it fails, the search knows that any
combination with a less sensitive parameter will also fail. The re-ordering of the
Hessian means that it is easy to skip the other combinations that are not allowed.
The next test to be applied is that of highest sensitivity. If the sensitivity of the
constructed combination is less than the current best candidate, the search moves on
the next combination. Again, by having the Hessian ordered, it can be quickly
determined whether the sum of all remaining parameters would be higher than the
highest sensitivity. If not, the search can skip the set of parameters that cannot form a
matrix with a high enough sensitivity to be considered.
Finally the correlations are tested. This is because the calculation of correlations
requires the Hessian to be inverted. Again as the correlations are calculated, as soon as
a single correlation is discovered that exceeds 90%, the search moves on to the next
combination.

If the construction of the correlations completes without discovering a correlation
greater than 90%, this combination must be the best one found yet and so it is stored
and the search continues.
This method is very fast and guaranteed to find the best matrix according to the
criteria:
1 Highest total sensitivity (determined as the sum along the leading diagonal of the
Hessian).
2 Lowest sensitivity is not less than 1% of the highest sensitivity (again sensitivity
refers to the terms along the leading diagonal of the Hessian).
3 No correlations more than 90%.
These criteria could be adjusted, though you must take care that the criteria you choose
define a weakly non-linear problem, otherwise the effort of analysis will not show any
benefit for the regression.

Wax and asphaltene precipitation in PVTi
Introduction
The solid precipitation model in PVTi handles waxes as a solid solution and
asphaltenes as an additional liquid phase.
The wax model assumes that paraffins and naphthalenes can precipitate as a solid
solution. In the asphaltene model it is assumed that the heaviest aromatic component
will form an asphaltenic liquid phase.
To use the solid precipitation model, the paraffin, naphthalene and aromatic fractions
for each carbon number are needed. These can be entered by hand, if known, or
generated from the correlations described in "The PNA distribution of heavy
components" on page 416 and "Critical properties of PNA species" on page 417.
The theory behind the wax model is explained in "Wax precipitation" on page 417.
The PNA distribution of heavy components

In PVTi, the components heavier than C6 may be split into their paraffinic (P),
naphthalenic (N) and aromatic (A) constituents. This is an important part of solid
precipitation as it is the P and N species, which generally form wax and the heaviest
aromatic is the major constituent (by weight) of the asphaltenic liquid phase.
The PNA distribution is then estimated as described by Nes and Westerns, see [Ref. 49].
v = 2.51 ( n – 1.4750 ) – ρ + 0.8510
w = ρ – 0.8510 – 1.11 ( n – 1.4750 )
3660
A% = 430v + ------------ ( for v < 0 )
MW
3660
A% = 670v + ------------ ( for v > 0 )
MW
10000
R = 820w + --------------- ( for w > 0 )
MW
10600
R = 1440w + --------------- ( for w < 0 )
MW
N% = R – A%
P% = 100 – R [EQ 8.214]
Where P%, N% and A% are the percentages of paraffinic, naphthalenic and aromatic
constituents respectively; ρ is the liquid density in gcm-3 at 20oC and 1 atm, and n is
the refractive index of the true boiling point (TBP) fraction.
n is given by Riazi and Daubert’s correlation, see [Ref. 50]:

n = ⎛ -------------
1 + 2I⎞
- [EQ 8.215]
⎝ 1–I⎠
where I is a characterization factor given by:

– 0.02269 0.9182
I = 0.3773T B SG
Hint If a full PNA specification of the sample is available, the mole fractions can be
used in the normal way, over-riding the default estimates from this splitting
procedure. First enter the single carbon number (SCN) fluid; perform the PNA
split, then type in the measured mole fractions.
Critical properties of PNA species

In PVTi the critical properties of the PNA species are set up using the correlations of
Riazi and Al-Sahhaf, see [Ref. 51], which are all of the form:
c
ln ( θ ∞ – θ ) = a – bMW [EQ 8.216]
where MW is the mole weight of the hydrocarbon and the constants a, b, c and θ ∞ have
been determined for various basic properties of PNA species (refer to the reference for
tables of their values).
Note The critical properties of the PNA species can be tuned just as for any other
components.
In addition to the usual set of critical properties, the melting points of the PNA species
are also determined, for use in the wax precipitation.
For paraffins, the melting point is given by the correlation of Won (1986), see [Ref. 52]:
f
T = 374.5 + 0.02617MW – 20172
--------------- [EQ 8.217]
MW
For naphthalenic and aromatic species, the melting point is given by (Pan, Firoozabadi
and Fotland, 1997 - see [Ref. 53]):
f – 0.00855MW
T = 333.45 – 419e [EQ 8.218]
where, in each equation, MW is the mole weight of the hydrocarbon.
Wax precipitation
In PVTi it is considered that the paraffinic and naphthalenic species of components
heavier than C15 can form waxes and that the heaviest aromatic component forms the
asphaltenic liquid phase.
Chemically, the waxes that drop out of hydrocarbon fluids at lower temperatures, are
known to contain paraffins and, to a lesser extent, naphthalenes.

In PVTi the method of Pedersen et al. (see [Ref. 56]) has been improved to use the
paraffin and naphthalene components, rather than a “wax forming component”. This
means that the critical properties correlations for the paraffins and naphthalenes, as
given by the correlations in "Critical properties of PNA species" on page 417, are used
in determining the solid precipitation.
The wax is modeled as a solid-solution and so we can apply an adjusted liquid fugacity
and solve for the wax phase using a full multiphase equation-of-state flash calculation.
50
For non-PNA species the solid fugacity is e , which effectively means that they
cannot form wax.
The adjustment to the liquid fugacity is given by:
f f f f
S L Δh i ⎛ T ⎞ Δc Pi ⎛ T ⎞ Δc Pi ⎛ T i⎞
f pure i ( P,T ) = f pure i ( P,T ) × exp --------- ⎜ 1 – -----i⎟ – ----------- ⎜ 1 – -----i⎟ – ----------- ln ⎜ -----⎟ [EQ 8.219]
RT ⎝ T⎠
f R ⎝ T⎠ R ⎝ T⎠
i
f
where Δh i is the enthalpy of fusion for component i; Δc Pi is its heat capacity of fusion;
f
and T i is its melting point temperature.
Enthalpy of fusion
The enthalpy of fusion for paraffins comes from the correlation by Won [Ref. 52]:
f f
Δh i = 0.1426MW i T i [EQ 8.220]
For napthalenes the correlation of Lira-Galeana et. al. (1996) [Ref. 57] was used:
f f
Δh i = 0.0527MW i T i [EQ 8.221]
For aromatics the correlation of Pan et al. [Ref. 53] was used:
f f
Δh i = 11.2T i [EQ 8.222]
Heat capacity of fusion

The heat capacity of fusion for all P, N, A species is given by the correlation of Pedersen
et al.[Ref. 58]:
–4
Δc Pi = 0.3033MW i – 4.635 × 10 MW i T [EQ 8.223]
The asphaltenic liquid phase

The asphaltenic phase is known to form at temperatures higher than the melting point
of the heavy aromatic molecule that is the solid asphaltene. Maximum precipitation
occurs close to the bubble point of the liquid and the precipitated phase is a thick, black
liquid, see [Ref. 59]. After filtering, an asphaltene deposit is recovered.

In PVTi there are a special set of default binary interaction coefficients introduced for
the interactions between the heaviest aromatic component and the light components.
This is based on the critical volumes of the components and takes a similar form to that
of the Cheuh-Prausnitz binary interactions :
⎛ 2V 1 ⁄ 6 V 1 ⁄ 6 ⎞ θ
k ij = A 1.0 – ⎜ ---------------------------------⎟
ci cj
[EQ 8.224]
⎜ 1⁄3 1 ⁄ 3⎟
⎝ V ci + V cj ⎠
where kij is the binary interaction coefficient between the ith and jth components; Vc is
the critical volume of the ith or jth component; A is the Cheuh-Prausnitz parameter,
which can be varied as a special regression variable, the value used for interactions
with the heaviest hydrocarbon is 10A; θ has been fitted using a number of asphaltene
data-sets and is 6.0 for interactions involving hydrocarbons with mole weights less
than that of SCN C7. It is zero for interactions involving hydrocarbons C7 and heavier,
except for the heaviest aromatic component for which it is 12.0.
This scheme is identical to the Cheuh-Prausnitz scheme for hydrocarbons up to C6,
when the PNA split is not made, the Cheuh-Prausnitz binaries are the same as in pre-
2002A versions of PVTi.
For P, N and all A species apart from the heaviest aromatic, the binary interactions are
zero. The strong interactions between the heaviest aromatic component and the light
components creates the conditions for a precipitating asphaltenic liquid phase, which
maximizes near the bubble line. Thus mimicking the behavior witnessed in laboratory
tests.
Discussion
PVTi uses a consistent, single fluid model for all calculations, so the critical properties
used in matching, say, the differential liberation experiment, are used in calculating the
wax appearance temperature. Consequently all the observations, including those of
the solid phase, can be regressed at once, leading to a more consistent and complete
single description for use in both reservoir and process simulations.

Cleaning samples contaminated with oil-based mud
Introduction
Oil-based muds are in widespread use and often contaminate PVT samples taken at the
well-site.
PVTi offers two methods for cleaning oil-based muds, "Removing oil-based mud
contamination by skimming" on page 420 describes a method that can be used when
the composition of the contaminant is not known. If the contaminant composition is
known, "Removing oil-based mud contamination by subtraction" on page 421 provides
an accurate method for removing the contaminant.
The methods used in PVTi are based on the work of Gozalpour et al, see [Ref. 54].

skimming
In naturally occurring hydrocarbon fluids, a semi-log straight-line relationship is seen
between increasing mole fraction and increasing mole weight for components C8 and
heavier, see [Ref. 57].
Many oil-based muds contain components in the range, C10 - C23, any contamination
of the reservoir fluid, therefore, appear as a positive deviation from the semi-log
straight-line behavior.
The assumptions made for the skimming method are: firstly, that the semi-log straight-
line behavior is exhibited by the fluid; and that there are uncontaminated components,
heavier than the heaviest contaminated component. The second assumption is
required for the straight-line to be constructed.
The method involves constructing the straight-line between the C8 component and the
uncontaminated heavy components. This line then gives the reservoir fluid and the
excess mole-fractions provide the composition of the contaminating oil-based mud.
In PVTi the light-end point is the first component with a mole weight heavier than 100
(if library components are used this is C8). The heavy end point is the lowest mole-
fraction before the plus-fraction.
Caution It is important that the sample being cleaned has heavier components than
the oil-based mud, so that there is an uncontaminated point to use in
constructing the straight line.
If the composition of the oil-based mud is known, it is better to use the subtraction
method described in "Removing oil-based mud contamination by subtraction" on page
421.

subtraction
As described in "Removing oil-based mud contamination by skimming" on page 420,
there is a semi-log straight-line relationship amongst the heavier components in a
naturally occurring hydrocarbon fluid that can be exploited to allow the removed of
oil-based mud contamination.
The subtraction method is applicable when the composition of the contaminant is
known.
In this case, a small amount of the contaminant is subtracted from the sample. The
resultant sample is tested for RMS fit to a semi-log straight-line defined as the same line
used in the skimming method.
A numerical optimization is performed to find the amount of contaminant that must
be removed to minimize this RMS fit.
Note The restriction that the oil-based mud contain components C10 - C23 does not
apply here. The mud can be of any composition.
Note The assumption that there are uncontaminated components, heavier than the
oil-based mud components, is not necessary for this method.
Hint If the composition is known, the subtraction method is a better choice than the
skimming method.

Mixing and recombination of samples
Introduction
PVTi offers both sample mixing and sample recombination. There is a subtle difference
between mixing and recombining. In mixing, you provide a mole fraction or gas-oil
ratio for the mix; these are used and the mixed sample is created directly. In
recombining, you enter a target gas-oil ratio for the mixture. In that case the proportion
of mixing is determined iteratively using the scheme outlined in "Recombination" on
page 422.
Mixing
Mixing is an addition of the two samples. If the mole fraction of the second sample is
specified, the mixture is a weighted sum. If gas-oil ratio (GOR) is specified, this value
is first translated to a mole fraction at the temperature and pressure requested.
The conversion from GOR to mole fraction (MF) is as follows:
n
Assuming n moles of Sample 2 are mixed with 1 mole of Sample 1, then F = ------------
1+n
RT std
The GOR is then given by GOR = ⎛ ------------⎞ --------------------- where R is the universal gas
F
⎝ 1 – F⎠ P V
std oil
constant, Tstd and Pstd are standard temperature and pressure respectively and Voil is
the oil volume at the temperature and pressure you specified. If you do not specify the
pressure, it is taken as the saturation pressure of the sample
Recombination
In recombination, the aim is to create a mixture with the stock tank gas-oil ratio
specified.
In this case, the temperature and pressure you enter are used as the first stage in a two-
stage separator. The second stage is always at standard conditions. The GOR you enter
is the target GOR for the mixture. This is used as an initial guess and the fluids are
mixed in the usual way (see "Mixing" on page 422). When this mixture is passed
through the separators, a stock-tank GOR is calculated. A regression is performed,
adjusting the mixing combination until the stock-tank GOR is equal to the target value
you requested.
The mix applied at each iteration is recorded in the log window.

Mixing and recombination of samples
ECLIPSE Thermal Export Module
Introduction
In PVTi and ECLIPSE Compositional we deal with an isothermal flash. This means
that, for a particular cell in a simulation, we know the composition of the fluid summed
over all the phases and the pressure and temperature. We try to minimize the Gibbs
Free Energy in order to determine how each component splits across the different
phases present. In effect we try to find the K-values, which are the unknown variables.
ECLIPSE Compositional assumes that the temperature of each cell stays constant over
time, even if there is a distribution of temperatures across the cells to start with. It is
assumed that if a particular drop of fluid moves from one cell with temperature, T1 to
another cell with temperature T2 that the fluid takes on the temperature T2. When the
Thermal option is used in ECLIPSE Compositional this is not the case. Thermal
diffusion is allowed to take place over time, which means that the temperature in each
cell is a free parameter that needs to be determined using a different type of flash; one
at constant energy. We therefore know the pressure, functional form of the K-Values
(K=K(P,T)), and the total energy of the system but the temperature is unknown.
The long-term goal of this ECLIPSE Thermal module is to use PVTi’s powerful
Equation of State (EoS) functionality to provide extensive support for the Thermal
option in the ECLIPSE Compositional simulator. The first step was to introduce
functionality into PVTi 2003A to calculate an optimal K-value functional relation. This
can then be used within ECLIPSE Thermal to perform these flashes at constant energy.
For PVTi 2004A the module has been extended to write out a whole file containing a
series of keywords that can be used as the PVT input for an ECLIPSE Thermal
simulation. In addition to the standard keywords written out for an ECLIPSE
Compositional run, for example acentric factors (ACF), critical temperatures (TCRIT)
etc., PVTi writes out additional keywords that enables ECLIPSE Thermal to calculate
accurate values for K-values oil/gas densities and oil/gas phase enthalpies.
• "K-Values" on page 423, "Oil Density" on page 425,
• "Gas phase density" on page 427,
• and "Enthalpy" on page 428 explain how the relevant keywords are used in
ECLIPSE Thermal to calculate the appropriate quantities.
For a brief non-technical summary of these keywords see "Outline of keywords for
ECLIPSE Thermal" on page 387.
K-Values
For PVTi 2004A you have the option of either using Crookston’s equation (the KVCR
keyword) or Wilson’s formula (the KVWI keyword).
Hint In fact Wilson’s formula is a simplified version of Crookston’s equation, which

is explained below.

The coefficients of Crookston’s equation are written out by PVTi using the KVCR
keyword. The general functional form of the K-Value correlation we use is given by
Crookston’s formula:
K ( P, T ) = ⎛⎝ A + --- + CP⎞⎠ exp ⎛⎝ -------------⎞⎠

B –D
[EQ 8.225]
P T –E
If a simplified version of equation [EQ 8.225] is used, given by setting A, C, E=0 that is:
K ( P, T ) = ⎛ ---⎞ exp ⎛ – D
----⎞
B
⎝ P⎠ ⎝ T⎠
[EQ 8.226]
then B and D can be calculated by using at least two experimentally determined

observations. Alternatively, Wilson’s formula can be used to estimate K-Values, which
is given by the formula:
Pc T
K ( P, T ) = ⎛ ------⎞ exp 5.372697 ( 1 + ω ) ⎛ 1 – ----c-⎞
⎝ P⎠ ⎝ T⎠
[EQ 8.227]
where P, T are the pressure and temperature and K , P c , T c and ω are the K-value,
critical pressure, critical temperature and acentric factor respectively for a particular
component. In fact Crookston’s equation is a generalized form of Wilson’s formula,
which can be obtained by setting:
A = 0
B = Pc
C = 0
T
D = 5.372697T ( 1 + ω ) ⎛ ----c- – 1⎞
⎝T ⎠
E = 0
in equation [EQ 8.225].
There are also tabulated values for B and D for certain components. The central
problem we face though is that K-values for a particular component are not functions
of just temperature and pressure, as Wilson’s formula would suggest, but also of the
types and quantities of other substances present. Consider the following. If we have
a fluid of say 50% methane and 50% decane then each component has its own K-value
at a particular temperature and pressure. If we now add, say toluene, to the mixture
then the K-values of methane and decane changes and we also have the K-value for the
new component. This illustrates that changing the composition of a fluid clearly effects
the K-values of respective components.
Also, for heavier hydrocarbon components (anything above about C7), the K-values
are an increasing monotonic function of pressure. It is clear that Wilson's formula is
completely inadequate here as K=Pc/P is clearly a decreasing monotonic function of
pressure. The full Crookston expression can cope with such components by using the
A and C terms.

PVTi can provide valuable assistance here because it provides its own experimental
data using the flash; as we know that the flash accurately reproduces experimental
observations. The ECLIPSE Thermal export facility then calculates the values of a
chosen set of Crookston coefficients so that the correlation best approximates a set of
flashes performed by PVTi in the pressure and temperature range you specified.
This correlation is unique to the fluid sample and specified pressure and temperature
range and ensures that the Thermal option in ECLIPSE Compositional is using
accurate approximations to component K-values.
Oil Density
PVTi writes out the keywords PREFT, TREFT, DREFT, CREF and THERMEX1, which
ECLIPSE Thermal uses internally to calculate the oil density of the fluid at a specified
pressure and temperature.
Algorithm
ECLIPSE Thermal calculates the molar oil density b o of the fluid using the following
set of equations:
b o is given by:
b o = 1 ⁄ V oil [EQ 8.228]
where the molar specific volume V oil of the oil phase is calculated using Amagat’s law
of partial volumes:
Nc
c c
V oil = ∑ x V oil [EQ 8.229]
c=1
c
and x is the mole fraction of component c in the oil phase. The component oil phase
c
volume V oil is given by:
c
V oil = MW
c
------------ [EQ 8.230]
c
ρ
c
where MW is the molecular weight of component c given by the MW keyword, and the
c
component oil phase density ρ is given by:
c
ρ
c ref
ρ = ---------------------------------------------------------------------------------------------
- [EQ 8.231]
( 1 + C T1 ( T – T ref ) ) ( 1 – C p ( P – P ref ) )

c
The reference density ρ is defined by ECLIPSE Thermal’s DREF keyword (or PVTi’s
ref
DREFT keyword); the standard temperature and pressure, T ref and P ref are defined
by keywords TREF (TREFT in PVTi) and PREF (PREFT in PVTi); the thermal expansion
coefficient C T1 is defined by the THERMEX1 keyword; and the component isothermal
compressibility C P is defined by the CREF keyword.
The quantities C T1 and C p are defined by the usual thermodynamic relations:
1 ∂V
C T1 = --- ------ [EQ 8.232]
V ∂T P = PREF
∂V
C p = – --1- ------ [EQ 8.233]
V ∂P T = TREF
Remarks
It is important to note that PVTi’s DREF, TREF keywords are, in general, not the same
as ECLIPSE Thermal’s DREF, TREF keywords.
PVTi’s DREF, TREF

In PVTi the reference density DREF is an observed/measured quantity of each
component at the reference temperature, TREF and standard pressure (14.7 psia). DREF
is then used in conjunction with PVTi’s predicted density at these conditions (using a
2-parameter equation of state) to calculate volume shifts for each component. Volume
shifts were introduced as a “third parameter” into EoS models to improve calculations
of liquid density. Therefore the standard values of TREF for each component in the
literature are chosen such that this component is in the liquid phase at the temperature
TREF and standard pressure.
ECLIPSE Thermal’s DREF, TREF, PREF

In ECLIPSE Thermal it can be seen from equation [EQ 8.230] that we require
knowledge of a reference density measured at a reference pressure and temperature.
The reference density is used along with the isothermal and isobaric expansion
coefficients (CREF and THERMEX1) to extrapolate linearly in order to obtain the density
of the oil given an arbitrary temperature and pressure. PVTi’s reference quantities are
not used because they may be very different to the conditions in the reservoir; and so
it would be unreasonable to expect accurate answers using a linear extrapolation.
In order to obtain reference parameters that are typical of the conditions in the
reservoir PVTi calculates the quantities TREFT, PREFT and DREFT. TREFT and PREFT
are the average of the maximum and minimum reservoir temperatures/pressures
respectively. You input these extreme reservoir values of temperature and pressure on
PVTi’s Export for ECLIPSE Thermal panel. DREFT for each component is then the
density calculated by PVTi’s EoS model at pressure PREFT and temperature TREFT.
Assuming that the component is in the liquid phase at these conditions then DREFT is
a good point to perform linear extrapolation using [EQ 8.230] to find the density of the
component at an arbitrary temperature and pressure in the reservoir.

However, if the component is not in the liquid phase (which is the case for lighter
components) at the initial values of TREFT and PREFT then we set TREFT=TREF
(PVTi’s TREF) and PREFT=14.7psia, which we know guarantees that the component is
in the liquid phase. We then use [EQ 8.230] to extrapolate to reservoir conditions and
calculate the properties of the component. These lighter components contribute a
relatively small amount to the overall density of the fluid and we have found this
approach to give accurate values of oil density.
The isothermal compressibility and thermal expansion coefficient are calculated by
PVTi using [EQ 8.231]and[EQ 8.232] with the appropriate values of TREFT and DREFT
TREFT, PREFT, DREFT are the same as ECLIPSE Thermal’s TREF, PREF, DREF. PVTi
uses alias names because, as explained, it has already has definitions for TREF and
DREF. ECLIPSE Thermal automatically recognizes these aliases and knows that it is
dealing with its own internal keywords.
Gas phase density

PVTi writes out the keywords TREFT, PREFT, DREFT and ZFACTOR, which ECLIPSE
Thermal uses internally to calculate the molar gas phase density of the fluid at a
specified pressure and temperature.
Algorithm
The molar density of the gas phase b g is given by:
b g = 1 ⁄ V gas [EQ 8.234]
where the molar specific volume V gas of the gas phase is
Nc
c c
V gas = ∑ y V gas [EQ 8.235]
c=1
c
assuming no water is present. y is the mole fraction of component c in the gas phase.
c
Each hydrocarbon component gas phase molar volume V gas is obtained from the gas
law,
c c
PV gas = Z RT [EQ 8.236]
c
where Z is specified with the ZFACTOR keyword.

Remarks
c
The algorithm outlined above assumes that Z does not change as a function of
c
temperature. Z is calculated by PVTi for each component by performing an EoS flash
on a fluid consisting purely of the relevant component. If a single gas phase is found
c
then Z is set to the value of the Z-factor returned by PVTi’s EoS code. If a liquid phase
c
is returned then Z is set to ECLIPSE Thermal’s default value of 0.96.
Enthalpy
• "Algorithm" on page 427 outlines the three different approaches ECLIPSE Thermal
can take in calculating oil and gas phase enthalpies.
• "Remarks" on page 428 then explains which of these approaches PVTi supports.
Algorithm
ECLIPSE Thermal calculates the enthalpy of the oil phase using a mole fraction
weighted average of the component enthalpies:
Nc
c c c
H oil = ∑ x ⋅ MW ⋅ H
oil
[EQ 8.237]
c=1
c c
where x is the mole fraction of component c in the oil phase and MW is the
component molecular weight given by the MW keyword.
The enthalpy of the hydrocarbon components in the gas phase is calculated using a
mole fraction weighted average of the component enthalpies:
Nc
c c c
H gas = ∑ y ⋅ MW ⋅ H gas [EQ 8.238]
c=1
c
where y is the mole fraction of component c in the gas phase.
The component enthalpies in the oil and gas phases are calculated from component
specific heats and heats of vaporization. Specific heat values must be given for each
hydrocarbon component in at least one fluid phase. If they are only specified in one
phase, then the heat of vaporization should be given. The following options are
therefore available for the fluid phase enthalpies:
1 Set the specific heat of components in the oil and gas phases and the heat of
vaporization at the standard temperature.
2 Set the specific heat of components in the oil phase and the heat of vaporization as
a function of temperature.
3 Set the specific heat of components in the gas phase and the heat of vaporization
as a function of temperature.

Solid phase enthalpies are only required in simulations where a solid phase is present.
Specific heats
c c
The specific heats C oil , C gas of a component c in the oil and gas phases are given by
c c c
C oil = C oil + C oil ⋅ ( T – T st )
1 2
c c c
C gas = C gas + C gas ⋅ ( T – T st )
1 2
where the oil phase coefficients are defined by keywords SPECHA and SPECHB and the
gas phase coefficients are defined by keywords SPECHG and SPECHH.
T st is the standard temperature defined by STCOND. The specific heats must always be
positive.
If oil phase specific heats are defined, the molar enthalpy of component c in the oil
phase is:
c c c 1 c 2
H oil = ∫ C oil dT = C oil ⋅ ( T – T st ) + --- ⋅ C oil ⋅ ( T – T st )
1 2 2
[EQ 8.239]
T st
c
For temperatures above the critical temperature, T > T crit , the oil phase molar enthalpy
is set equal to the gas phase molar enthalpy.
If gas phase specific heats are defined, the molar enthalpy of component c in the gas
phase is
c c c
H gas = H vaps + ∫ C gas dT [EQ 8.240]
T st
c c 1 c 2
= H vaps + C gas ⋅ ( T – T st ) + --- ⋅ C gas ⋅ ( T – T st )
1 2 2
c
where H vaps is the heat of vaporization at the standard temperature, defined by
HEATVAPS.
If a solid phase is simulated, the molar enthalpy of component c in the solid phase is
c c c 1 c 2
H sol = ∫ C sol dT = C sol ⋅ ( T – T st ) + --- ⋅ C sol ⋅ ( T – T st )
1 2 2
[EQ 8.241]
T st
Heat of vaporization
c
A temperature dependent heat of vaporization ΔH ( T ) can be specified instead of
either the oil or gas specific heat.
If the specific heat of a component c in the oil phase is not specified, then the oil phase
enthalpy is calculated from the gas component enthalpy by

c c c
H oil = H gas – ΔH ( T ) [EQ 8.242]
Alternatively, if the specific heat of a component c in the gas phase is not specified,
then the gas phase enthalpy is calculated from the oil component enthalpy
c c c
H gas = H oil + ΔH (T) [EQ 8.243]
The heat of vaporization is given by [Ref. 57]:
c c B T < T crit
c
ΔH ( T ) = A ⋅ ( 1 – T ⁄ T crit ) [EQ 8.244]
c c
ΔH ( T ) = 0 T ≥ T crit [EQ 8.245]
where the constant A is defined by the HEATVAP keyword, the exponent B is defined
c
by the HEATVAPE keyword, and T crit is the component critical temperature defined
by the TCRIT keyword.
Heats of vaporization are usually obtained at the normal boiling point T nb . The
constant A , defined by the HEATVAP keyword can be found by inverting :
c
ΔH ( T nb )
A = -----------------------------------------
- [EQ 8.246]
B
( 1 – T nb ⁄ T crit )
c
where ΔH ( T nb ) is the heat of vaporization at the normal boiling point. The exponent
B for each component, entered with keyword HEATVAPE, is usually set to a value in
the range 0<B<=1.
Remarks
PVTi uses the oil and gas phase component specific heat coefficients (SPECHA, SPECHB,
SPECHG and SPECHH) and the heats of vaporization at the standard temperature
(HEATVAPS) for each component. As explained in the previous section, the oil and gas
phase enthalpies are then calculated using [EQ 8.239] and [EQ 8.240].
In the future this functionality could be extended so that you could specify in PVTi
which of the three options to calculate enthalpies you wish to take advantage of in
ECLIPSE Thermal. PVTi would then export the relevant keywords depending on
which option you had selected.

Units
Appendix A
Units
General information
This option allows the current unit and temperature conventions to be changed. The
options available are:
• Metric unit system
• Field units
• Laboratory units
• PVT-metric units
The temperature conventions are:
• Degrees Kelvin
• Degrees Celsius
• Degrees Rankine
• Degrees Fahrenheit
Any of the unit conventions are compatible with any of the temperature options. The
unit conventions may be changed at any point in a session, existing quantities being
converted to the new units. PVT-metric units are the same as metric, except in that
pressure is measured in atmospheres. In addition, it is possible to enter mole
compositions as fractions (of unity) or percentages (up to 100%), and pressures can be
specified in absolute or gauge units.
Units 431
The units for each data quantity are given in the table below.
Table A.1 Units
Quantity Metric Field Lab PVT-M

Length metres feet cms metres
Depth
Time days days hours days
Density 3 3 gm ⁄ cc 3
kg ⁄ m lb ⁄ ft kg ⁄ m
Pressure barsa psia atma atma
(absolute)
Pressure difference bars psi atm atm
Compressibility 1 ⁄ bars 1 ⁄ psi 1 ⁄ atm 1 ⁄ atm
Viscosity cPoise cPoise cPoise cPoise
Permeability mDarcy mDarcy mDarcy mDarcy
Liquid surface volume 3 stb scc 3
sm sm
Gas surface volume 3 mscf scc 3
sm sm
Reservoir volume 3 rb rcc 3
rm rm
Liquid surface volume 3 stb ⁄ day scc ⁄ hour 3
rates m ⁄ day sm ⁄ day
Gas surface volume 3 mscf ⁄ day scc ⁄ hour 3
rate sm ⁄ day sm ⁄ day
Reservoir volume 3 rb ⁄ day rcc ⁄ hour 3
(rate) rm ⁄ day rm ⁄ day
Formation volume 3 3 rb ⁄ mscf rcc ⁄ scc 3 3
factor (liquid) rm ⁄ sm rm ⁄ sm
Gas-oil ratio 3 3 mscf ⁄ stb scc ⁄ scc 3 3
sm ⁄ sm sm ⁄ sm
Oil-gas ratio 3 3 stb ⁄ mscf scc ⁄ scc 3 3
sm ⁄ sm sm ⁄ sm
Volume 3 3 cc 3
m ft m
Specific volume 3 -ml 3 -ml -ml 3 -ml
m ⁄ kg ft ⁄ lb cc ⁄ gm m ⁄ kg
Energy kJ btu J kJ
Enthalpy -ml -ml -ml -ml
kJ ⁄ kg btu ⁄ lb J ⁄ gm kJ ⁄ kg
Specific heat – ml – ml – ml – ml
kJ ⁄ kg ⁄K btu ⁄ lb ⁄R J ⁄ gm ⁄K J ⁄ kg ⁄K
Thermal conductivity kJ ⁄ m ⁄ d ⁄ K btu ⁄ ft ⁄ d ⁄ R J ⁄ cm ⁄ h ⁄ K kJ ⁄ m ⁄ d ⁄ K
Gas-oil ratio (oil at 3 3 mscf ⁄ rb scc ⁄ rcc 3 3
reservoir) sm ⁄ rm sm ⁄ rm
432 Units
Table A.1 Units (Continued)

Gauge pressure barsg psig atmosg atmosg
Surface tension dyne ⁄ cm dyne ⁄ cm dyne ⁄ cm dyne ⁄ cm
Transmissibility 3 cPrb ⁄ d ⁄ psi cPcc ⁄ h ⁄ a 3
cPm ⁄ d ⁄ bar cPm ⁄ d ⁄ a
A number of constants that depend upon the unit convention are used. These are:
Table A.2 Constants

Gravity constant 0.000098066 0.00694444 0.000967841 0.0000967841
Darcy constant 0.00852702 0.00112712 3.6 0.00864
Atmos. pressure 1.01325 14.6959 1.0 1.0
Density of air 1.2232 0.076362 0.0012232 1.2232
Gas constant R 0.083143 10.732 82.055776 0.08205576
Standard conditions are taken as one atmosphere and 60 °F .
Some useful conversion factors

The following table contains conversion factors.
Table A.3 Conversion factors
Quantity Conversion factor

Length 1m = 3.28084 ft
1ft = 0.30480 m
Volume 3 3 = 6.2898110 bbl
1m = 35.3146600 ft
3 3 = 0.1781076 bbl
1ft = 0.02831685 m
Mass 1kg = 2.20462300 lb
1lb = 0.45359237 kg
Density 3 3 = 1.0E-3 gm ⁄ cc
1kg ⁄ m = 0.06242797 lb ⁄ ft
3 3 = 0.01601846 gm ⁄ cc
1lb ⁄ ft = 16.0184600 kg ⁄ m
Pressure 1 bar = 14.5037700 psi = 0.986923 atm
1 psi = 0.06894757 bar = 0.068046 atm
Units 433
Table A.3 Conversion factors (Continued)
Quantity Conversion factor

Gas-oil 3 3 = 5.614583E-3 mcf ⁄ bbl
1m ⁄ m
1 mcf ⁄ bbl 3 3
= 178.1076000 m ⁄ m
Gravities API (141.5/liq.grav.)-131.5
Temperatures T (deg K ) = T (deg R )/1.8
Conversion of gas moles to volumes at Z=1

The volume occupied by one mole of gas, with an ideal gas Z -factor of unity, at
temperature T (deg R ) and pressure p (psia), is:
Vmolar = RT ⁄ p = 379.39445 at p = 14.70 psi, T = 519.67 deg R
Vmolar = RT ⁄ p = 23.650203 at p = 1.013 bar, T = 288.15 deg K
The number of moles in unit volume V of gas with an “ideal gas” Z -factor of unity, at
temperature T and pressure p , is:
lb-moles = p ⁄ RT = 0.002635779 at p = 14.70 psi, T = 519.67 deg R
kg-moles = p ⁄ RT = 0.042282930 at p = 1.1013 bar, T = 288.15 deg K
434 Units
Symbols
Appendix B
Symbols
f iL Fugacity of component i in the liquid phase
f iV Fugacity of component i in the vapor phase
Ki Equilibrium constant of component i
L Mole fraction of liquid phase
m1 Constant in Martins equation of state.
m2 Constant in Martins equation of state.
p Pressure of mixture.
p ci Critical pressure of component i .
p ri p
Reduced pressure of component i : p ri = ------i- .
p ci
T Temperature of mixture.
Tc Critical temperature of component i .
T ri T
Reduced temperature of component i : T ri = ------i- .
T ci
V Mole fraction of vapor phase.
xi Mole fraction of component i in the liquid phase.
Symbols 435
yi Mole fraction of component i in the vapor phase.
Z Z -factor in equation of state.

δ ij Binary interaction coefficient.
ΩA Omega- A values for each component.

i
ΩB Omega- B values for each component.

i
ωi Acentric factor for component i .
436 Symbols
Bibliography
Wilson, G. “A Modified Redlich-Kwong Equation of State, Application to General Physical Data

Calculations”. [Ref. 1]
Paper No. 15C Ohio, presented at the AIChE 65th National Meeting, Cleveland. May 4-7,
(1969).
Martin, J.J. “Cubic Equations of State - Which?” [Ref. 2]
IEC Fundamentals, Vol. 18, Page 81 May, (1979).

Coats, H. “Simulation of Gas Condensate Reservoir Performance”, [Ref. 3]
SPE paper no. 10512, presented at the Sixth SPE Symposium on Reservoir Simulation, New
Orleans January 31st - February 3rd, (1982).
Whitson, C.H. “Topics on: Phase Behaviour and Flow of Petroleum Reservoir Fluids”, [Ref. 4]
Ph.D. Thesis, The University of Trondheim, Norwegian Institute of Technology,

Department of Petroleum and Chemical Engineering, August, (1983).
Pedersen, S.K., “Viscosity of Crude Oils”, [Ref. 5]
Fredenslund, Aa., Chemical Engineering Science, Vol. 39, Page 1011-1016 No. 6, (1984)
Christensen, P.L., and
Thomassen, P.
Whitson, C.H., and “Evaluating Constant-Volume Depletion Data”, [Ref. 6]
Torp, S.B. J. Pet. Tech. Page 610-620. March, (1983)
Lohrenz, J., “Calculating Viscosity of Reservoir Fluids from their Composition”, [Ref. 7]
Bray, B.G., and J. Pet. Tech. Page 1171 (1964); Trans., AIME,. 231.
Clark, C.R.
Crowe, C.M., and “Convergence Promotion in the Simulation of Chemical Processes - The General Dominant
Nishio, M. Eigenvalue Method”, [Ref. 8]
AIChEJ, Vol. 23, No. 3 Page 528-529. May (1975)

Dennis, J.E., and “Numerical Methods for Unconstrained Optimisation and Non-linear Equations”, [Ref. 9]
Schnabel, R.B. Prentice-Hall Inc., New Jersey, (1983)
Bibliography 437
Kesler, M.G., and “Improved Predictions of Enthalpy of Fractions”, [Ref. 10]
Lee, B.I. Hydro. Proc. Page 153-158 March (1976)
Cavett, R.H. “Physical Data for Distillation Calculations - Vapor-Liquid Calculations”, [Ref. 11]
Proc. 27th API Meeting, San Francisco, Page 351-366,(1962)

Riazi, M.R., and “Simplify Property Predictions”, [Ref. 12]
Daubert, T.E. Hydro. Proc. Page 115-116 March (1980)
Lee, B.I., and “Improved Vapor Pressure Predictions”, [Ref. 13]
Kesler, M.G. Hydro. Proc. Page 163-167. July (1980)
Edmister, W.C. “Applied Hydrocarbon Thermodynamics”, [Ref. 14]
Part 4: Compressibility Factors and Equation of State, Pet. Refiner Page 173-179 April,
(1958)
Michelsen, M.L. “The Isothermal Flash Problem. Part I. Stability”, [Ref. 15]
Fluid Phase Equilibria, Vol. 9 Page 1-19, (1982)

Katz, D.L. et al. “Handbook of Natural Gas Engineering”, [Ref. 16]
McGraw-Hill Company Page 129, (1982)

Reid, R.C, “The Properties of Gases and Liquids”, [Ref. 17]
Prausnitz, J.M., and McGraw-Hill Company, Third Edition, (1977)
Sherwood, T.K.
Bashbush, J.L. “A Method to determine K-values from Laboratory Data and its Applications”, [Ref. 18]
SPE 10127 (1981)

Hoffmann, A.E., Equilibrium Constants for a Gas-Condensate System”, [Ref. 19]
Crump, J.S., and Trans. AIME., 198 Page 1-10.
Hocott, C.R.
Alani, G.H., and “Volumes of Liquid Hydrocarbons at High Temperatures and Pressures”, [Ref. 20]
Kennedy, H.T. Page Trans. AIME, 219, Page 288-292 (1960)
Dranchuk, P.M., “Computer Calculations of Natural Gas Compressibility Factors using the Standing and Katz
Purvis, R.A., and Correlation”, [Ref. 21]
Robinson, D.B. Institute of Petroleum Technical Series, No. IP 74-008, (1974).
Wichert, E., and “Calculate Zs for Sour Gases”, [Ref. 22]
Aziz, K. Hydrocarbon Processing, 51, Page 119 May (1972)
Lee, A.L., “The Viscosity of Natural Gases”, [Ref. 23]
Gonzalez, M.H., and J. Pet. Tech. Page 997-1000 August, (1966)
Eakin, B.E.
Standing, M.B. “A Set of Equations for Computing Equilibrium Ratios of a Crude Oil/Natural Gas System
at pressures below 1000 psia”, [Ref. 24]
J. Pet. Tech. Page 1193-1195 September (1979)

Peneloux, A., “A Consistent Correction for Soave-Redlich-Kwong Volumes”, [Ref. 25]
Rauzy, E., and Fluid Phase Equilibria 8 Page 7-23, (1982)
Freze, F.
Søreide, I., “Procedures for Reservoir Fluid Characterisation using an Equation of State Model”,[Ref. 26]
Reffstrup, J., and Report RE 88-8, Laboratory for Energetics, Danish Tech. Univ., Lyngby.
Whitson, C.H.
438 Bibliography
Coats, K.H., and “Application of a Regression-Based EOS PVT Program to Laboratory Data”, [Ref. 27]
Smart, G.T. SPE Res. Eng. Page 277-299 May (1986)
Herning, F., and “Calculation of the Viscosity of Technical Gas Mixtures from the Viscosity of the Individual
Zippener, L. Gases”, [Ref. 28]
Gas u. Wasserfach 79Page 69-73, (1936)

Stiel, L.I., and Thodos, “The Viscosity of Nonpolar Gases at Normal Pressures”, [Ref. 29]
G. AIChEJ Page 611-615 (1961)
Pedersen, K.S., “Properties of Oils and Natural Gases”, [Ref. 30]
Fredenslund, Aa., and Gulf Publishing Company, Houston (1989)
Thomassen, P.
Abramowitz, M., and Handbook of Mathematical Functions [Ref. 31]
Stegun, I.A. New York Page 923, (1972)
Whitson, C.H., “ C 7+ Characterisation of Related Equilibrium Fluid using the Gamma Distribution”,[Ref. 32]
Anderson, T.F., and Paper in C 7+ Characterisation, Edited by Mansoori G.A., and Chorn L.G., Taylor and
Søreide, I. Francis, New York, (1989).
McCain, W.D. “The Properties of Petroleum Fluids”, [Ref. 33]
PennWell Publishing Company, Second Edition, Tulsa (1990)

Søreide, I. “Improved Phase Behaviour Predictions of Petroleum Reservoir Fluids from a Cubic Equation
of State”, [Ref. 34]
Ph.D. Thesis, The University of Trondheim, Norwegian Institute of Technology,

Department of Petroleum and Chemical Engineering April (1989)
Søreide, I., and “Peng-Robinson Predictions for Hydrocarbons, CO 2 ,N 2 , and H 2 S With Pure Water and
Whitson, C.H. NaCl -Brines”, [Ref. 35]
Fluid Phase Equilibria, Vol. 77 Page 217-240, (1992)

Pedersen, S.K., and “Chemical Engineering Science”, [Ref. 36]
Fredenslund, Aa. Vol. 42, No. 1 Page 182-186, (1987)
Reudelhuber, F.O., “A Compositional Material Balance Method for Prediction of Recovery from Volatile Oil
and Hinds, R.F. Depletion Drive Reservoirs”, [Ref. 37]
Petroleum Transactions, AIME, Vol. 210 Page 19, (1957)

Jensen, F., and “Calculation of First Contact and Multiple Contact Minimum Miscibility Pressures”,[Ref. 38]
Michelsen, M.L. In Situ, 14(1) Page 1-17, (1990)
Gmehling, J., “Vapor-Liquid Equilibrium Data Collection DECHEMA Chemistry Data Series”, [Ref. 39]
Onken, U., and DECHEMA, Frankfurt/Main, (1981)
Arlt, W.
Fragor Reference Manual Vol. 2.4, Chapter 5 [Ref. 40]
J. Pet. Tech. [Ref. 41]
Page 720-723 July 1986.

Pedersen K.S., “Characterisation of Gas Condensate Mixtures”, [Ref. 42]
Thomassen P, and Adv. Thermodynamics 1 Page 137, (1989b)
Fredenslund Aa
Winn F.W. “Physical Properties by Monogram”, [Ref. 43]
Pet. Refiner 36 Page 157-159, (1957)
Bibliography 439
Sim W.J. and “Prediction of vapor-Liquid Equilibrium of Undefined Mixtures”, [Ref. 44]
Daubert T.E. Ind. Eng. Chem. Process. Des. Dev. 19 Page 386-393, (1980)
Joergensen, M. and “Optimisationofpseudo-componentselectionforcompositionalstudiesofreservoirfluids”,[Ref. 45]
Stenby, E.H SPE 30789, 70th Annual SPE Technical conference & exhibition. Dallas TX (1995)
Pedersen, K. Unpublished [Ref. 46]
1997
Michelsen, M.L. and “Calculation of critical points from cubic 2-constant equation of state”, [Ref. 47]
Heidemann, R.A. AIChE Journal (Vol 27 No3) Page 769, (1980)
Heidemann, R.A. and “The Calculation of Critical Points”, [Ref. 48]
Khalil, A.M. AIChe Journal 26 No 5 Page 521, (1981)
Nes, K. and Westerns, “Aspects of the Constitution of Mineral Oils”, [Ref. 49]
H. A. van Elsevier, New York, (1951)
Riazi, M. R. and “Prediction of the composition of petroleum fractions”, [Ref. 50]
Daubert, T. E. Ind. Eng. Chem. Process Des. Dev., Vol. 19, Page 289 -294, (1980)
Riazi, M. R. and Al- “Physical properties of n-alkanes and n-alkylhydrocarbons: Application to petroleum
Sahhaf, T. A. mixtures”, [Ref. 51]
Ind. Eng. Chem. Res., Vol. 34, Page 4145 -4148, (1995)
Won [Ref. 52]
Fluid Phase Equilibria, Vol. 30, Page 265, (1986)

Pan, H., Firoozbadi, A. “Pressure and composition effect on was precipitation: Experiment data and model
and Fotland, P., results”, [Ref. 53]
SPE Prod. and Fac - SPE 36740, Page 250, (1997)

Gozalpour, F., Densh, “Predicting reservoir fluid phase and volumetric behaviour from samples contaminated with
A., Tehranio, D. H., oil-based mud”, [Ref. 54]
Todd, A.C. and Tohidi, SPE 56747, (1999)
B.,
Katz, D. “Overview of Phase Behavior in Oil and Gas Production”, [Ref. 55]
JPT, Page 1205-1214, (1983)

Pedersen, K. S. “Prediction of cloud point temperatures and amount of wax precipitation”, [Ref. 56]
SPE 27629, (1995)

Lira-Galeana, C., “Thermodynamics of wax precipitation in petroleum mixtures”, [Ref. 57]
Firoozbadi, A., and AIChE J., Vol. 42, Page 239, (1995)
Prausnitz, J. M.
Pedersen, K. S., “Wax precipitation from North Sea crude oils. 4. Thermodynamic Modeling”, [Ref. 58]
Skovborg, P. and Energy and Fuels, Vol. 5 (6), Page 924, (1991)
Ronningsen, H. P.
Rydahl, A. K., “Modeling of live oil asphaltene precipitation”, [Ref. 59]
Pedersen, K. S., and
Hjermstad, H. P.
“Properties of Oils and Natural Gases”, [Ref. 60]
Chapter 10, ISBN 0-87201-588-2

Ahmed, T. SPE 18532 [Ref. 61]
Liu, K. SPE 66363 [Ref. 62]
440 Bibliography
Trebble, M.A. “A preliminary evaluation of two and three phase flash initiation procedures”, [Ref. 63]
Fluid Phase Equilibria. 53, Page 113-122, (1989)

Aasberg-Petersen, K. “The Viscosity of Hydrocarbon Fluids”, [Ref. 64]
Fluid Phase Equilibria. 70, Page 293-308(1991)

Teja, A.S. & Rice, P. “Generalised Corresponding States Method for the Viscosities of Liquid Mixtures”, [Ref. 65]
“Ind. Eng. Chem. Fundam., 20,Page 77(1981)
Hanley, H.J.M., “Equation of the Viscosity and Thermal Conductivity Coefficients of Methane”, [Ref. 66]
McCarty, R.D. & Cryogenics, 15, Page 413,(1975)
Haynes, W.M.
McCarthy, R.D., “A Modified Benedict-Webb-Rubin Equation of State for Methane Using Recent
Experimental Data” [Ref. 67]
Cryogenics, 14, Page 276 (1974)
Bibliography 441
442 Bibliography
Index
keywords . . . . . . . . . . . . . . . . 171 COMB

A Boiling point temperatures . . . . . 278 keyword . . . . . . . . . . . . . . . . . . 189
Bubble point pressure. . . . . . . . . . 356 COMB section
Acentric factors. . . . . . . . . . . . 115, 179 keywords . . . . . . . . . . . . . . . . . 170
ACF keyword . . . . . . . . . . . . . . . . . 179 COMBINE keyword. . . . . . . . . . . . 190
ACHEUH keyword . . . . . . . . . . . . 180 Command line . . . . . . . . . . . . . . . . . 31
A-coefficient . . . . . . . . . . . . . . . . . . 180 C Components . . . . . . . . . . . . . . . . . . 240
Alani-Kennedy liquid densities . . 324 Calorific values . . . . . . . . . . . 158, 185 Compositional
Alkanes . . . . . . . . . . . . . . . . . . . . . . 308 CALVAL keyword. . . . . . . . . . . . . 185 Blackoil. . . . . . . . . . . . . . . . . . . 394
Aromatics . . . . . . . . . . . . . . . . . . . . 309 grading with depth . . . . . . . . 394
Cascade . . . . . . . . . . . . . . . . . . . . . . 161
Asphaltene and Wax System . . . . . 77 Condensate systems . . . . . . . . . . . . 358
Cavett . . . . . . . . . . . . . . . .114, 186, 313
Consistency tests. . . . . . . . . . . . . . . 320
CCE . . . . . . . . . . . . . . . . . . . . . . . . . 359
correlations . . . . . . . . . . . . . . . 403
Cell compositions . . . . . . . . . . . . . 321
Constant Mole Fraction116, 314 to 315
B Cell volumes . . . . . . . . . . . . . . . . . 322
Conversion factors . . . . . . . . . . . . . 433
Celsius. . . . . . . . . . . . . . . . . . . . . . . 431
BIC keyword . . . . . . . . . . . . . . . . . . 182 CORRACF keyword. . . . . . . . . . . . 191
CHAR components . . . . . . . . . . . . 312
Binary Interaction CORRCP keyword . . . . . . . . . . . . . 192
Coefficients . . . 115, 182, 333, 353 CHARACT keyword. . . . . . . . . . . 186
Correlation. . . . . . . . . . . . . . . . . . . . 192
Coefficients default . . . . . . . . 197 Characterization . . . . . . . . . . . . . . 186
Critical pressures . . . . . . . . . . . . . . 253
Black oil tables Cheuh-Prausnitz . . . . . . . . . . 158, 180
Critical temperatures . . . . . . . . . . . 279
differential . . . . . . . . . . . . . . . . 199 CMF. . . . . . . . . . . . . . . . . . . . . . . . . 315
Critical viscosity . . . . . . . . . . . . . . . 347
BLACKOIL CNAMES keyword . . . . . . . . . . . . 187
keyword. . . . . . . . . . . . . . . . . . 184 Critical volumes . . . . . . . . . . . . . . . 289
CO2 Rich Fluids. . . . . . . . . . . . . . . . 78
Blackoil Critical Z-factors . . . . . . . . . 301 to 302
Coats . . . . . . . . . . . . . . . . . . . . . 29, 188
Compositional simulation . . 394 CVD . . . . . . . . . . . . . . . . . . . . . . . . . 359
COATS keyword . . . . . . . . . . . . . . 188
BLACKOIL section Cycloparaffins . . . . . . . . . . . . . . . . . 309
Index 443
D F I
Data analysis. . . . . . . . . . . . . . . . . . . 85 Fahrenheit. . . . . . . . . . . . . . . . . . . . 431 Individual observation weights . . 245
DBG file . . . . . . . . . . . . . . . . . . 103, 196 Field units. . . . . . . . . . . . . . . . . . . . 431 Internal library . . . . . . . . . . . . . . . . 227
Dead oil tables . . . . . . . . 181, 194, 226 FIT keyword. . . . . . . . . . . . . . . . . . .211 Isomers . . . . . . . . . . . . . . . . . . . . . . . 308
DEADGAS keyword . . . . . . . 193, 297 Fluid definition . . . . . . . . . . . . . . . . 27
DEADOIL keyword . . . . 181, 194, 226 Fourth-degree polynomial . . . . . . 346
DEBUE keyword . . . . . . . . . . . . . . 195 FRAC keyword . . . . . . . . . . . . . . . 212 K
Debug. . . . . . . . . . . . . . . . . . . . . . . . 161 FRAGOR keyword . . . . . . . . . . . . 213
Katz-Firoozabadi . . . . . . . . . . . . . . 158
DEBUG keyword . . . . . . . . . . . . . . 196 Fragor method . . . . . . . . . . . . . . . . 213
Kelvin . . . . . . . . . . . . . . . . . . . . . . . . 431
Default Binary Interaction Fugacities . . . . . . . . . . . . . . . . . . . . 330
Coefficients . . . . . . . . . . . 197 Kesler-Lee . . . . . . 114 to 115, 186, 313
FVFREF keyword . . . . . . . . . . . . . 214
DEFBIC keyword . . . . . . . . . . . . . . 197 Keyword errors . . . . . . . . . . . . . . . . 174
DEGREES keyword . . . . . . . . . . . . 198 Keywords
PVTi . . . . . . . . . . . . . . . . . . . . . 177
Dew point . . . . . . . . . . . . . . . . . . . . 358
Differential black oil tables . . . . . . 199
G K-value checks . . . . . . . . . . . . . . . . 404
Gas formation volume factor. . . . 356 KVTABLE keyword . . . . . . . . . . . . 222
DREF keyword . . . . . . . . . . . . . . . . 200
Dry gas tables . . . . . . . . . . . . . 193, 297 Gas gravity (density) . . . . . . . . . . 356
Gas viscosity . . . . . . . . . . . . . 327, 356
Gas-oil ratio . . . . . . . . . . . . . . . . . . 356 L
definition . . . . . . . . . . . . . . . . 159
E GeoQuest Program Launcher . . . . 31
Laboratory experiments . . . . . . . . 355
Laboratory units . . . . . . . . . . . . . . . 431
ECHO keyword . . . . . . . . . . . . . . . 202 GI
keyword . . . . . . . . . . . . . . . . . 215 LBC keyword . . . . . . . . . . . . . . . . . 224
ECLIPSE 300 . . . . . . . . . . . . . . . . . . 204
nodes . . . . . . . . . . . . . . . . . . . . 215 LBC viscosity correlation. . . . . . . . 160
Edmister . . . . . . . . . . . . . 115, 186, 313
GOR LBCCOEF keyword . . . . . . . . . . . . 225
EoS
definition . . . . . . . . . . . . . . . . 159 Lee-Kesler . . . . . . . . . . . . . . . . . . . . 313
omega values . . . . . . . . . . . . . 247
regression to measured data. . 29 GRAF. . . . . . . . . . . . . . . . . . . . . . . . 159 LIB components . . . . . . . . . . . . . . . 312
EOS keyword . . . . . . . . . . . . . . . . . 203 GRBYALL keyword . . . . . . . . . . . 217 Library components . . . . . . . . . . . . 105
EOSOUT keyword . . . . . . . . . . . . . 204 GRBYMIX keyword . . . . . . . . . . . 218 Linear thermal expansion coefficient
Equation of state. . . . . . . . . . . 203, 330 GRBYSAM keyword . . . . . . . . . . . 219 336
fitting to experimental results 53 GROUP keyword . . . . . . . . . . . . . 216 Liquid formation volume factors . 214
Equilibrium K-values . . . . . . . . . . 323 GROUP section . . . . . . . . . . . . . . . 169 Liquid saturation . . . . . . . . . . . . . . 157
EXP keyword . . . . . . . . . . . . . . . . . 205 keywords . . . . . . . . . . . . . . . . 169 LNAMES keyword. . . . . . . . . 187, 227
Experiments . . . . . . . . . . . . . . . . . . 205 Grouping . . . . . . . . . . . . . . . . .118, 190 LOG files . . . . . . . . . . . . . . . . . . . . . 103
components . . . . . . . . . . . . . . 216
Export . . . . . . . . . . . . . . . . . . . . . . . 148 Lohrenz-Bray-Clark . . 224 to 225, 255
Equilibration . . . . . . . . . . . . . . 147 GRPBYWGT keyword . . . . . . . . . 220
Gas Reservoir . . . . . . . . . . . . . 147
Oil Reservoir . . . . . . . . . . . . . . 146
VFP . . . . . . . . . . . . . . . . . . . . . . 149 M
Export modules . . . . . . . . . . . . . . . 144 H
Martin. . . . . . . . . . . . . . . . . . . . . . . . 332
Hoffman-Crump plot . . . . . . . . . . . 89
Material balance checks . . . . . . . . . . 28
HYDRO keyword . . . . . . . . . . . . . 221
Maximum step size . . . . . . . . . . . . 229
Hydrocarbon . . . . . . . . . .113, 221, 308
MAXIT keyword. . . . . . . . . . . . . . . 228
444 Index
MAXSTEP keyword. . . . . . . . . . . . 229 OBSIND keyword . . . . . . . . . . . . . 245 PVO file . . . . . . . . . . . . . . . . . . . . . . 103
MDP keyword . . . . . . . . . . . . . . . . 230 Oil density . . . . . . . . . . . . . . . . . . . 356 PVP file. . . . . . . . . . . . . . . . . . . . . . . 103
MES file . . . . . . . . . . . . . . . . . . . . . . 103 Oil formation volume factor . . . . 356 PVT data for ECLIPSE simulators . 29
MESSAGE keyword. . . . . . . . . . . . 231 Oil viscosity . . . . . . . . . . . . . . . . . . 356 PVTLIB annexe . . . . . . . . . . . . . . . . 195
Metric units . . . . . . . . . . . . . . . . . . . 431 OMEGAA, OMEGAB keywords. 247 PVT-metric units. . . . . . . . . . . . . . . 431
Michelsen’s stability criterion . . . 331 On-line help . . . . . . . . . . . . . . . . . . 161
MINDELP keyword. . . . . . . . . . . . 232 Opening a project . . . . . . . . . . . . . 105
MINSTEP keyword . . . . . . . . . . . . 233 OPTIONS keyword . . . . . . . . 210, 248 Q
MIX keyword . . . . . . . . . . . . . . . . . 234 OUTECL3 keyword . . . . . . . . . . . 250
Quality control . . . . . . . . . . . . . . . . . 85
Mixing . . . . . . . . . . . . . . . . . . . . . . . 234 OUTECL3 section . . . . . . . . . . . . . 172
Mnemonics . . . . . . . . . . . . . . . . . . . 187 keywords . . . . . . . . . . . . . . . . 172
Modified form of Peng-Robinson 256

Modified Whitson . . . . . . . . . . . . . 314 R
Modified Whitson Splitting . . . . . 317 P Rankine . . . . . . . . . . . . . . . . . . . . . . 431
MODSPEC keyword . . . . . . . . . . . 235 Recovery calculations . . . . . . . . . . 405
PARACHOR keyword . . . . . . . . . 252
MODSYS keyword. . . . . . . . . . . . . 236 RECOVERY keyword . . . . . . . . . . 258
Paraffins . . . . . . . . . . . . . . . . . . . . . 308
Molar volume correction. . . . . . . . 335 Redlich-Kwong . . . . . . . . . . . . . 27, 331
PCRIT keyword . . . . . . . . . . . . . . . 253
Mole fractions . . . . . . . . . . . . . . . . . 303 Reference densities . . . . . . . . . . . . . 200
PEARCE keyword . . . . . . . . . . . . . 254
Mole weight of the plus fraction . 239 Reference temperatures . . . . . . . . . 283
Pedersen . . . . . . . . . . . . . .115, 186, 313
Molecular weights . . . . . . . . . . . . . 238 REGRESS section . . . . . . . . . . . . . . 170
PEDERSON keyword . . . . . . . . . . 255
Monitor option . . . . . . . . . . . . . . . . 161 keywords . . . . . . . . . . . . . . . . . 170
Peneloux . . . . . . . . . . . . . . . . . . . . . 335
MOSES keyword . . . . . . . . . . . . . . 237 Regression
Peneloux et al. volume shift. . . . . . 27
Moses method . . . . . . . . . . . . . . . . 237 operation . . . . . . . . . . . . . . . . . 211
Peng-Robinson. . . . . . . . . . . . . 27, 331 target. . . . . . . . . . . . . . . . . . . . . 260
Multi-feed Split. . . . . . . . . . . . . . . . 116 modified form of . . . . . . . . . . 256 variables . . . . . . . . . . . . . . . . . . 285
MW keyword . . . . . . . . . . . . . . . . . 238 Phase diagrams . . . . . . . . . . . . . . . 312 REGTARG keyword . . . . . . . . . . . . 260
MWS keyword . . . . . . . . . . . . . . . . 239 Plus fraction splitting . . . . . . . . . . 212 Reservoir temperature . . . . . . . . . . 261
PRCORR keyword . . . . . . . . . . . . 256 Reudelhuber and Hinds . . . . . . . . 258
Pressure nodes Riazi-Daubert . . . . . . . . . 114, 186, 313
N automatic insertion . . . . . . . . 232
RTEMP keyword . . . . . . . . . . . . . . 261
Probability density model . . . . . . 314
Napthenes . . . . . . . . . . . . . . . . . . . . 309 RUNSPEC . . . . . . . . . . . . . . . . . . . . 166
Program options . . . . . . . . . . 156, 248
NCOMPS keyword . . . . . . . . . . . . 240 RUNSPEC keyword . . . . . . . . . . . . 262
PSEUCOMP keyword. . . . . . . . . . 257
NEW. . . . . . . . . . . . . . . . . . . . . . . . . 103 RUNSPEC section
PSEUCOMP section . . . . . . . . . . . 171 keywords . . . . . . . . . . . . . . . . . 167
NEW files . . . . . . . . . . . . . . . . . . . . 103 keywords . . . . . . . . . . . . . . . . 171
NEWPVI keyword . . . . . . . 241 to 242 Pseudo-compositional tables . . . . 257
Newton-Raphson . . . . . . . . . . . . . . 331 Pseudo-critical
NOECHO keyword . . . . . . . . . . . . 243 temperatures, pressures . . . . 326 S
Non-hydrocarbon. . . . . . . . . . 113, 221 Pseudoise SALINITY keyword . . . . . . . . . . . . 263
definition . . . . . . . . . . . . . . . . . 28
Sample composition. . . . . . . . . . . . 265
Pseudoised
SAMPLES keyword . . . . . . 264 to 266
EoS data . . . . . . . . . . . . . . . . . . 29
O Pure . . . . . . . . . . . . . . . . . . . . . . . . . 308
SAMTITLE keyword . . . . . . . . . . . 267
Saturated black oil reservoirs 393, 396
OBS keyword . . . . . . . . . . . . . . . . . 244 Pure components. . . . . . . . . . . . . . 310
Saturation
Observations . . . . . . . . . . . . . . . . . . 244 PVI file . . . . . . . . . . . . . . . . . . . . . . 103
liquid . . . . . . . . . . . . . . . . . . . . 157
Index 445
SAVCOMP keyword . . . . . . . . . . . 268 Temperature convention . . . . . . . 198 Volatile oil reservoirs . . . . . . . 393, 397
Saving compositions . . . . . . . . . . . 268 Thermal expansion . . . . . . . . . . . . 158 Volume expansion coefficient . . . . 280
Schmidt-Wenzel . . . . . . . . . . . . . . . . 27 THERMX keyword . . . . . . . . . . . . 280 Volume shift
SCN . . . . . . . . . . . . . . . . . . . . . . . . . 114 Thomassen . . . . . . . . . . . .115, 186, 313 corrections . . . . . . . . . . . . . . . . 158
SCN groups. . . . . . . . . . . . . . . . . . . 310 Tiled. . . . . . . . . . . . . . . . . . . . . . . . . 161 Volume shift parameters
dimensionless . . . . . . . . . . . . . 274
SCT keyword . . . . . . . . . . . . . . . . . 269 TITLE keyword . . . . . . . . . . . . . . . 281
Volume translation . . . . . . . . . . . . . 335
SCT split . . . . . . . . . . . . . . . . . . . . . 269 TLOW keyword. . . . . . . . . . . . . . . 282
Semi-Continuous Thermodynamics . ToolTips. . . . . . . . . . . . . . . . . . . . . . 162
116, 314 TREF keyword . . . . . . . . . . . . . . . . 283
split. . . . . . . . . . . . . . . . . . . . . . 269
True-boiling-point distillation . . . 313 W
Separator liquid volumes . . . . . . . 158
Trust region radius . . . . . . . . 229, 233 WAT100 keyword . . . . . . . . . . . . . . 293
SG keyword . . . . . . . . . . . . . . . . . . 270
Two-parameter equation of state 335 WAT200 keyword . . . . . . . . . . . . . . 294
Shift parameters . . . . . . . . . . . . . . . 336
WAT300 keyword . . . . . . . . . . . . . . 295
SIMULATE . . . . . . . . . . . . . . . . . . . 271
Water properties
SIMULATE keyword . . . . . . . . . . . 271
SIMULATE section U ECLIPSE 100 . . . . . . . . . . . . . . 293
ECLIPSE 300 . . . . . . . . . . . . . . 295
keywords . . . . . . . . . . . . . . . . . 170 Undersaturated reservoirs . . . . . . 393 ECLIPSE GI option . . . . . . . . . 294
Simulation of experiments . . . . . . . 28 Unit convention . . . . . . . . . . . . . . . 284 VFPi . . . . . . . . . . . . . . . . . . . . . 296
Single Carbon Number . . . . . . . . . 114 Unit types . . . . . . . . . . . . . . . . . . . . 155 Watson characterization factor . . . 313
Soave-Redlich-Kwong . . . . . . . 27, 331 UNITS keyword. . . . . . . . . . . . . . . 284 WATVFP keyword . . . . . . . . . . . . . 296
Søreide . . . . . . . . . . . . . . . . . . . . . . . 336 UNIX . . . . . . . . . . . . . . . . . . . . . . . . . 31 WHIT keyword . . . . . . . . . . . . . . . . 298
SPECHA,B,C,D keyword . . . . . . . 272 UNIX platforms . . . . . . . . . . . . . . . . 31 Whitson . . . . . . . . . . . . . . 116, 314, 316
Specific heat . . . . . . . . . . . . . . . . . . 158 probability density function . 230
User components. . . . . . . . . . . . . . 312
splitting . . . . . . . . . . . . . . . . . . 298
Specific heat capacity coefficients 272
Whitson and Torp. . . . . . . . . . . . . . 299
SPLIT keyword. . . . . . . . . . . . . . . . 273
WHITSON keyword . . . . . . . . . . . 299
SPLIT section
keywords . . . . . . . . . . . . . . . . . 169 V Whitson-Torp . . . . . . . . . . . . . . 29, 188
Splitting . . . . . . . . . . . . . . . . . . . . . . 116 Vapor Z-factor . . . . . . . . . . . . . . . . 325 Window . . . . . . . . . . . . . . . . . . . . . . 161
SSHIFT keyword . . . . . . . . . . . . . . 274 VAR keyword. . . . . . . . . . . . . . . . . 285 Windows platforms . . . . . . . . . . . . . 31
Standard conditions. . . . . . . . . . . . 275 VCRIT keyword. . . . . . . . . . . . . . . 289 Winn . . . . . . . . . . . . . . . . . 114, 186, 313
Standard pressure . . . . . . . . . . . . . 156 VCRITVIS keyword . . . . . . . . . . . 290
Standard temperature . . . . . . . . . . 156 VEC file . . . . . . . . . . . . . . . . . . . . . . 103
STCOND keyword. . . . . . . . . . . . . 275 VERSION keyword . . . . . . . . . . . . 291 X
Surface tension . . . . . . . . . . . . . . . . 334 VFP . . . . . . . . . . . . . . . . . . . . . . . . . 149
XMFVP keyword . . . . . . . . . . . . . . 300
Swelling test . . . . . . . . . . . . . . . . . . 361 section . . . . . . . . . . . . . . . . . . . 172
SYSTEM . . . . . . . . . . . . . . . . . . . . . . 166 VFP keyword . . . . . . . . . . . . . . . . . 292
SYSTEM keyword . . . . . . . . . . . . . 276 VFP section
SYSTEM section
keywords . . . . . . . . . . . . . . . . 172 Y
keywords . . . . . . . . . . . . . . . . . 168 VFP Tables . . . . . . . . . . . . . . . . . . . 386
YMFVP keyword . . . . . . . . . . . . . . 300
VFPi . . . . . . . . . . . . . . . . . . . . . 144, 282
View menu . . . . . . . . . . . . . . . . . . . 105
T Viewing the K-value Fits . . . . . . . . 83

Viscosities . . . . . . . . . . . . . . . . . . . . 347
Z
TBOIL keyword . . . . . . . . . . . . . . . 278 Viscosity calculations . . . . . . . . . . 404 ZCRIT keyword . . . . . . . . . . . . . . . 301
TCRIT keyword . . . . . . . . . . . . . . . 279
446 Index
ZCRITVIS keyword . . . . . . . . . . . . 302 ZMFVD keyword . . . . . . . . . . . . . 304
Z-factors, critical. . . . . . . . . 301 to 302 Zudkevitch-Joffe . . . . . . . 27, 331, 334
ZI keyword . . . . . . . . . . . . . . . . . . . 303
Index 447
448 Index

PVTi REFERENCE MANUAL PDF

Uploaded by

Copyright:

Available Formats

PVTi REFERENCE MANUAL PDF

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

PVTi REFERENCE MANUAL PDF

Uploaded by

Copyright:

Available Formats

PVTi

Service mark information

Chapter 1 - New developments...................................................................................................... 11

Chapter 2 - The Most Asked Questions About PVTi.................................................................... 15

Chapter 3 - Introduction ................................................................................................................. 27

Chapter 4 - Getting started ............................................................................................................ 31

Chapter 5 - Tutorials ....................................................................................................................... 33

Chapter 6 - Reference section ....................................................................................................... 99

Chapter 7 - Keywords ................................................................................................................... 177

Appendix A - Units........................................................................................................................ 431

Appendix B - Symbols.................................................................................................................. 435

Chapter 2 - The Most Asked Questions About PVTi.................................................................... 15

Chapter 3 - Introduction ................................................................................................................. 27

Chapter 4 - Getting started ............................................................................................................ 31

Chapter 5 - Tutorials ....................................................................................................................... 33

Chapter 6 - Reference section ....................................................................................................... 99

Chapter 7 - Keywords ................................................................................................................... 177

Chapter 8 - Technical Description ............................................................................................... 305

Appendix A - Units........................................................................................................................ 431

Appendix B - Symbols.................................................................................................................. 435

Chapter 2 - The Most Asked Questions About PVTi ................................................................... 15

Chapter 3 - Introduction ................................................................................................................. 27

Chapter 4 - Getting started ............................................................................................................ 31

Chapter 5 - Tutorials ....................................................................................................................... 33

Chapter 6 - Reference section ....................................................................................................... 99

Chapter 7 - Keywords ................................................................................................................... 177

Chapter 8 - Technical Description ............................................................................................... 305

Appendix A - Units........................................................................................................................ 431

Appendix B - Symbols.................................................................................................................. 435

Developments for 2005A

ECLIPSE Thermal Export facility

Export for API Tracking option in ECLIPSE

The Most Asked Questions About PVTi 15

What is PVTi used for? Why do we need it?

Where do I start? How do I set up a project within

16 The Most Asked Questions About PVTi

Creating Other Fluid Samples

How do I create an experiment along with a series

The Most Asked Questions About PVTi 17

Note Currently defined observations for an experiment can be edited in the

For more information on creating and editing experiments/observations see

What are the data limitations in PVTi?

18 The Most Asked Questions About PVTi

What is the Fluid Properties Estimation facility in

How do I perform regression on multiple fluid

The Most Asked Questions About PVTi 19

What is the difference between normal regression,

20 The Most Asked Questions About PVTi

What regression parameters should I choose?

• P c, T c and ω of any non-library component.

The Most Asked Questions About PVTi 21

How does PVTi support gas condensate

22 The Most Asked Questions About PVTi

Can PVTi be used if you know the composition of

You have observations

The Most Asked Questions About PVTi 23

How do I generate the asphaltene phase envelope

Caution You cannot plot theoretical predictions from PVTi of an experimental

24 The Most Asked Questions About PVTi