LBNL-6871E

USER’S MANUAL OF THE

TOUGH+ CORE CODE v1.5:
A GENERAL-PURPOSE SIMULATOR OF
NON-ISOTHERMAL FLOW AND
TRANSPORT THROUGH POROUS AND
FRACTURED MEDIA

G.J. Moridis and K. Pruess

Earth Sciences Division,

Lawrence Berkeley National Laboratory,
Berkeley, CA 94720

August 2014

This work was supported by the Assistant Secretary for Fossil Energy, Office of Natural Gas and
Petroleum Technology, through the National Energy Technology Laboratory, under the U.S.
Department of Energy, Contract No DE-AC02-05CH11231.
 PAGE LEFT INTENTIONALLY BLANK

ii
User’s Manual of the TOUGH+ Core Code v1.5:
A General-Purpose Simulator of
Non-Isothermal Flow And Transport
Through Porous And Fractured Media

G.J. Moridis and K. Pruess

Earth Sciences Division, Lawrence Berkeley National Laboratory
University of California, Berkeley, California

Abstract

TOUGH+ v1.5 is a numerical code for the simulation of multi-phase, multi-

component flow and transport of mass and heat through porous and fractured media, and
represents the third update of the code since its first release [Moridis et al., 2008].
TOUGH+ is a successor to the TOUGH2 [Pruess et al., 1991; 2012] family of codes for
multi-component, multiphase fluid and heat flow developed at the Lawrence Berkeley
National Laboratory. It is written in standard FORTRAN 95/2003, and can be run on any
computational platform (workstations, PC, Macintosh).
TOUGH+ v1.5 employs dynamic memory allocation, thus minimizing storage
requirements. It has a completely modular structure, follows the tenets of Object-Oriented
Programming (OOP), and involves the advanced features of FORTRAN 95/2003, i.e.,
modules, derived data types, the use of pointers, lists and trees, data encapsulation, defined
operators and assignments, operator extension and overloading, use of generic procedures,
and maximum use of the powerful intrinsic vector and matrix processing operations.
This report presents the core TOUGH+ v1.5 code, i.e., the part of the code that is
common to all its applications. It provides a description of the underlying physics and
thermodynamics of non-isothermal flow, of the mathematical and numerical approaches, as
well as a detailed explanation of the general (common to all applications) input
requirements, options, capabilities and output specifications. The core code cannot run by
itself: it needs to be coupled with the code for the specific TOUGH+ application option that
describes a particular type of problem. The additional input requirements specific to a
particular TOUGH+ application options and related illustrative examples can be found in
the corresponding User’s Manual.

iii
 PAGE LEFT INTENTIONALLY BLANK

iv
 TABLE OF CONTENTS
Abstract .................................................................................................................................................... iii  

List of Figures ........................................................................................................................................... ix  

List of Tables ............................................................................................................................................. xi  

1.0. INTRODUCTION ............................................................................................. 1  

1.1.   Background ..................................................................................................................................... 1  

1.2.   The TOUGH+ v1.5 Core Code ...................................................................................................... 3  

2.0 CONCEPTS, UNDERLYING PHYSICS, AND GOVERNING EQUATIONS .... 7  

2.1.   Modeled Processes and Underlying Assumptions........................................................................ 7  

2.2.   The Mass and Energy Balance Equation ...................................................................................... 9  

2.3.   Mass Accumulation Terms ............................................................................................................ 9  

2.4.   Heat Accumulation Terms ........................................................................................................... 10  

2.5.   Flux Terms ..................................................................................................................................... 12  

2.6.   Source and Sink Terms ................................................................................................................ 18  

2.7.   Thermophysical Properties .......................................................................................................... 19  

2.7.1.   Water ........................................................................................................................................... 19  
2.7.2.   Real Gases .................................................................................................................................. 20  
2.7.3.   Other Phases and Components ................................................................................................... 21  

2.8.   Porosity and Intrinsic Permeability Changes ........................................................................... 22  

2.8.1.   Changes caused by P- and T-dependence ................................................................................... 22  
2.8.1.   Changes caused by the evolution of solid phases ........................................................................ 23  

2.9.   Multiphase Diffusion .................................................................................................................... 27  

2.9.1.   General Considerations ............................................................................................................... 27  
2.9.2.   Diffusion Formulation in TOUGH+ ............................................................................................ 28  

2.10.   Wettability-Related Phenomena ............................................................................................... 30  

2.10.1.   Relative permeavility ................................................................................................................. 31  
2.10.2.   Capillary pressure ..................................................................................................................... 31  
2.10.3.   Effect of solid phase deposition on capillary pressure .............................................................. 34  
2.10.4.   Effect of solid phase deposition on relative permeability .......................................................... 34  
2.10.5.   Pore compressibility of unconsolidated media in the presence of cementing solid phases ...... 34  

2.11.   Description of Flow in Fractured Media .................................................................................. 45  

v
3.0. DESIGN AND IMPLEMENTATION OF TOUGH+ V1.5 ................................ 49  
3.1.   Primary Variables ......................................................................................................................... 49  

3.2.   Space and Time Discretization .................................................................................................... 51  

3.3.   The Newton-Raphson Iteration ................................................................................................... 53  

3.4.   Implications of the Space Discretization Approach................................................................... 55  

3.5.   Space Discretization of Diffusive Fluxes ..................................................................................... 55  

3.6.   Code Units of the TOUGH+ v1.5 Code ....................................................................................... 57  

4.0. INPUT DATA REQUIREMENTS ................................................................... 63  

4.1.   Input Procedure ............................................................................................................................ 64  
4.1.1.   Data Block/Keyword TITLE ....................................................................................................... 64  
4.1.2.   Keyword/Record ENDCY ........................................................................................................... 64  
4.1.3.   Keyword/Record ENDFI............................................................................................................. 65  
4.1.4.   Structure of TOUGH+ Standard Input Files .............................................................................. 65  

5.0. MEMORY SPECIFICATION AND ALLOCATION ........................................ 71  

5.1.   Data Block MEMORY ...................................................................................................................... 71  

5.2.   Internal Checks ............................................................................................................................. 78  

6.0. PHYSICAL PROPERTIES OF SYSTEM ...................................................... 81  

6.1.   Data Block ROCKS or MEDIA ....................................................................................................... 81  

6.3.   Data Block RPCAP ........................................................................................................................ 87  

6.3.1.   Two-Phase Relative Permeability Functions .............................................................................. 88  
6.3.2.   Two-Phase Capillary Pressure Functions .................................................................................. 91  

6.4.   Data Block DIFFUSION ............................................................................................................... 94  

6.4.1.   User Options for Multiphase Diffusion ...................................................................................... 98  

6.5.   Block-by-Block Permeability Modification ................................................................................ 99  

7.0. GEOMETRICAL REPRESENTATION, DOMAIN DISCRETIZATION,

AND GRID GENERATION .................................................................................. 101  
7.1.   TOUGH+ Convention for Geometrical Data ........................................................................... 101  

7.2.   Data Block ELEME ...................................................................................................................... 103  

7.3.   Data Block CONNE ...................................................................................................................... 106  

vi
 7.4.   The MeshMaker.f95 Facility .................................................................................................. 108  
7.4.1.   Inputs Related to Problem Definition and Dimensioning ......................................................... 108  
7.4.2.   Inputs Related to Domain Heterogeneity.................................................................................. 110  
7.4.3.   Inputs Related to Description of Boundaries ............................................................................ 113  
7.4.4.   Inputs for Grid Construction .................................................................................................... 116  

8.0. INITIAL CONDITIONS AND BOUNDARY CONDITIONS .......................... 127  

8.1.   Data Block INCON ...................................................................................................................... 127  

8.2.   Data Block INDOM ...................................................................................................................... 130  

8.3.   Data Block EXT-INCON ............................................................................................................. 131  

8.3.1.   Data Block GEOMETRY ........................................................................................................... 131  
8.3.2.   Data Block SEQUENCE ........................................................................................................... 132  
8.3.3.   Data Block LIST ........................................................................................................................ 133  
8.3.4.   Remaining Data Blocks in EXT-INCON ................................................................................... 134  
8.3.5.   Data Block COLUMN ............................................................................................................... 135  

8.4.   Implementing Initial Conditions ............................................................................................... 136  

8.5.   Implementing Boundary Conditions ......................................................................................... 138  

8.5.1.   General ..................................................................................................................................... 138  
8.5.2.   Data Block BOUNDARIES ....................................................................................................... 140  

9.0. SOURCES AND SINKS .............................................................................. 145  

9.1.   Data Block GENER ...................................................................................................................... 145  

9.2.   Discussion on sinks and sources ................................................................................................ 149  

10. COMPUTATIONAL PARAMETERS ............................................................ 151  

10.1.   Data Block PARAM .................................................................................................................... 151  

10.2.   Modification of Computational Parameters During the Course of a TOUGH+

Simulation ................................................................................................................................. 162  

10.3.   Data Block SOLVR .................................................................................................................... 166  

10.4.   Discussion on Linear Equation Solvers .................................................................................. 168  

11. OUTPUT SPECIFICATIONS ........................................................................ 171  

11.1   Output of Primary and Secondary Variables ........................................................................ 172  

11.2.   Data Block TIMES .................................................................................................................... 172  

11.3.   Data Block SUBDOMAINS ........................................................................................................ 173  

11.4.   Data Block INTERFACES ........................................................................................................ 178  

vii
 11.5.   Data Block SS_GROUPS ........................................................................................................... 187  

11.8.   Warning Output and Error Messages .................................................................................... 189  

Acknowledgements ................................................................................................................................ 191  

References ............................................................................................................................................... 193  

viii
LIST OF FIGURES

Figure 2.1. Schematic of pore channels, showing convergent-divergent geometry with a

succession of pore throats and pore bodies............................................................. 25

Figure 2.2. Tubes-in-series model of pore channels .................................................................. 26

Figure 2.3. Liquid and gas relative permeabilities based on the van Genuchten model
[Finsterle, 1999]........................................................................................................ 31

Figure 2.4. Schematic of probability density function p(r) for pore size distribution ................... 33

Figure 2.5. Compressibility of an unconsolidated porous medium impregnated with

cementing solid phases (ice and/or hydrates). In this example, SSmin = 0.15,
-8 -1 -9 -1
SSmax = 0.4, αPU = 10 Pa , αPL = 10 Pa and δ = 0.015.. ..................................... 43

Figure 2.6. Effect of the varying compressibility described in Figure 2.5 on the porosity of
an unconsolidated porous medium undergoing depressurization for various
levels of saturation SS of cementing solid phases. ................................................... 44

Figure 2.7. Idealized double porosity model of a fractured porous medium. .............................. 45

Figure 2.8. Subgridding in the method of "multiple interacting continua" (MINC)....................... 47

Figure 2.9. Flow connections in the “dual permeability” model. Global flow occurs
between both fracture (F) and matrix (M) grid blocks. In addition there is F-M
interporosity flow. ..................................................................................................... 48

Figure 3.1. Space discretization and geometry data in the integral finite difference method
(from Pruess et al. [1999]).. ...................................................................................... 53

Figure 4.1. The DIFFUSION data block, with examples of the

Diffusion_Key_Parameters and
Component_Diffusivities_in_Phases namelists. ......................................... 98

Figure 7.1. An example of a MeshMaker.f95 input file for the creation of a Cartesian 3D
grid. Note that no heterogeneous regions or boundaries are defined in this
grid. ........................................................................................................................ 121

Figure 7.2. An example of a MeshMaker.f95 input file for the creation of a single-layer
(1D) cylindrical grid. Note that no heterogeneous regions or boundaries are
defined in this grid. ................................................................................................. 121

Figure 7.3. An example of a MeshMaker.f95 input file for the creation of a large
Cartesian 3D grid with heterogeneous regions and defined boundaries................ 122

Figure 7.4. An example of a MeshMaker.f95 input file for the creation of a large
cylindrical 2D grid with multiple layers, heterogeneous regions and defined
boundaries.............................................................................................................. 123

Figure 8.1. An example of the NAMELIST-described termination data printed at the end of
the SAVE file from a TOUGH+ v1.5 simulation. ...................................................... 129

ix
Figure 8.2. An example of the namelist structure of the BOUNDARIES data block. The
time-variable data are provided and read in tabular form....................................... 129

Figure 10.1. An example of a Parameter_Update_File for parameter updating in the

course of a TOUGH+ simulation. ........................................................................... 164

Figure 10.2. An example of a Parameter_Update_File indicating three completed

parameter updates, in addition to another one (at the top of the file) that has
not yet been executed. ........................................................................................... 165

Figure 11.1. Examples of the SUBDOMAINS data block for tracking the evolution of volume-
averaged properties and conditions in specified subdomains. This data block
uses namelist-based formats for data inputs.......................................................... 178

Figure 11.2. Examples of the INTERFACES data block for tracking flows at specified
interfaces. This data block uses namelist-based formats for data inputs. ............. 186

Figure 11.3. Example of the SS_GROUPS data block for tracking flows through specified
groups of sources/sinks. This data block uses namelist-based formats for
data inputs. ............................................................................................................. 189

x
LIST OF TABLES

Table 4.1. TOUGH+ input data blocks. ........................................................................................ 68

xi
 PAGE LEFT INTENTIONALLY BLANK

xii
 1.0. Introduction

1.1. Background
TOUGH+ v1.5 is a numerical code for the simulation of multi-phase, multi-component

flow and transport of mass and heat through porous and fractured media, and represents

the third update of the code since its first release in a version focusing on the analysis of

system behavior of hydrate-bearing sediments [Moridis et al., 2008]. TOUGH+ is a

successor to the TOUGH2 [Pruess et al., 1999; 2012] family of codes for multi-

component, multiphase fluid and heat flow developed at the Lawrence Berkeley National

Laboratory (LBNL). It is written in standard FORTRAN 95/2003, and can be run on any

computational platform (workstations, PC, Macintosh).

TOUGH+ v1.5 has a completely modular architecture. Any member of the

TOUGH+ family of codes comprises three components: (a) the core TOUGH+ code that

is common to all applications related to the study of non-isothermal processes of flow and

transport through geologic media, (b) the code that is unique to a particular type of

1
application/problem (e.g., the properties and flow of a crude oil, the flow of water and air

through geologic media, etc.), and (c) supplemental TOUGH+ code units that describe

special physics and processes that are encountered in particular types of problems (e.g.,

code units that describe real gas properties, non-Darcian flow processes, salinity effects on

the properties of water, etc.) and are used by more than one application options.

Thus, the core TOUGH+ code – which is distributed as a separate entity by

LBNL– cannot conduct any simulations by itself, but needs additional units of

supplemental and problem-specific code before it can become operational. The additional

code solves the equation of state (EOS) corresponding to the specific problem; it is called

an application option or simply an option in the TOUGH+ nomenclature and is

distributed as a separate entity/product by LBNL. The term option – rather the older term

module or EOS that were used in the TOUGH2 [Pruess et al., 1999; 2012] nomenclature

– is used to avoid confusion, as the word module has a particular meaning in the

FORTRAN 95/2003 language of TOUGH+.

For example, to solve the problem of flow of water/brine and real gas flow through

a tight fractured porous medium (e.g., in a fractured shale reservoir), the TOUGH+ v1.5

code must be coupled with the REALGASBRINE v1.0 option [Moridis and Freeman,

2014]. A discussion of the processes that are particular to this problem, as well as a

detailed explanation of the additional (over those needed by the core code) input data

requirements, a description of the output options, and appropriate illustrative examples

with sample input and output files are included in the User’s Manual of the corresponding

application option.

2
1.2. The TOUGH+ v1.5 Core Code
While the underlying principles, physics and thermodynamics are similar to those used in

the TOUGH2 family of codes [Pruess et al., 1999; 2012], the code in TOUGH+ v1.5 is a

radical departure from the earlier language and architecture of TOUGH2. The FORTRAN

95/2003 language of TOUGH+ has enabled a drastically different architecture. It employs

dynamic memory allocation, thus minimizing storage requirements. It follows the tenets

of Object-Oriented Programming (OOP), and involves entirely new data structures that

describe the objects upon which the code is based. The basic objects are defined through

derived data types, and their properties and processes are described in modules and sub-

modules, i.e., entities that incorporate the object attributes and parameters in addition to

the procedures (corresponding to the older concepts of “functions” and “subroutines” in

FORTRAN 77) that describe its behavior and processes. As discussed earlier, the

TOUGH+ v1.5 code has a completely modular structure that is designed for maximum

traceability and ease of expansion.

By using the capabilities of the FORTRAN95/2003 language, the new OOP

architecture involves the use of pointers, lists and trees, data encapsulation, defined

operators and assignments, operator extension and overloading, use of generic procedures,

and maximum use of the powerful intrinsic vector and matrix processing operations

(available in the extended mathematical library of FORTRAN 95/2003). This leads to

increased computational efficiency, while allowing seamless applicability of the code to

multi-processor parallel computing platforms. The result is a code that is transparent and

compact, and frees the developer from the tedium of tracking the disparate attributes that

define the objects, thus enabling a quantum jump in the complexity of problem that can be

3
tackled. An additional feature of the FORTRAN 95/2003 language of TOUGH+ is the

near complete interoperability with C/C++, which allows the interchangeable use of

procedures written in either FORTRAN 95/2003 or C/C++, makes possible the seamless

coupling with external packages (such as the geomechanical commercial code FLAC3D

[Itasca, 2002] or ROCMECH [Kim and Morids, 2013]) and interaction with pre- and post-

processing graphical environments.

Note that TOUGH+ v1.5 still uses a large number of the inputs (and the input

formats) used by the conventional TOUGH2 code to fulfill the functional requirement

(part of the code design) of backward compatibility of the TOUGH+ family codes with

older input data files used in TOUGH2 [Pruess et al., 1999; 2012] simulations. However,

more advanced input data structures and formats are introduced in this version to support

and describe capabilities unavailable in earlier code versions. More powerful input data

file structures will be available in planned future releases of TOUGH+.

By solving the coupled equations of mass and heat balance, TOUGH+ v1.5 – coupled

with the appropriate application option – can model the non-isothermal phase behavior

and flow of fluids and heat in complex geologic media. The code can simulate problems

covering the range from laboratory- to field-scale. The only limitations on the size of the

domain to be simulated are imposed by the underlying physics and by the limitations of

the computing platform. Thus, if the volume of the domain and its subdivisions are such

that (a) a representative volume can be defined and (b) the fluid flow through the

porous/fractured media can be adequately described mathematically, then TOUGH+ can

be used for the solution of the problem if the appropriate option is available.

4
 This report on the core code of TOUGH+ v1.5 provides a detailed description of

the underlying equations of mass and energy balance in general multi-phase, multi-

component systems involving porous and fractured porous media. Flow through such

media is described by using multi-phase versions of Darcy’s law, with appropriate

extensions to describe non-linear behavior in gas flow through low-permeability media

[Klinkenberg, 1901]. The report also discusses the source on information for the

description of the thermophysical properties of commonly used fluids, such as water, real

gases, etc., and includes additional subjects that are common to a wide range flow and

transport processes such as multi-phase, multi-component diffusion, the pressure and

temperature dependence of porosity (and the corresponding effect on permeability), gas

solubility in water, flow through fractured media, and wettability-related issues (relative

permeability and capillary pressure) in deformable media, including the effect of solid

phases such as precipitating salts or the formation of phases such as ice and gas hydrates.

The spatial and temporal discretization of the integrodifferential equations of mass

and energy balance equations results in a set of strongly nonlinear (in general) algebraic

equations that need to be satisfied in every subdivision (element) of the discretized

domain. These fully implicit equations are linearized using the Newton-Raphson iteration,

and the resulting Jacobian matrix equation is solved – using one of the several solutions

options, including a family of preconditioned conjugate gradient solvers and a direct

solver – until an acceptable convergence of the solution at each timestep is attained. The

accurate solution of the mass and energy balance equations being the basic method

employed in TOUGH+ v1.5 (and in all members of the TOUGH2 family of codes [Pruess

et al., 1999; 2012]), the timestep control is always subjugated to the requirement for an

5
accurate solution and is a process driven mainly by the progress of the simulation rather

than by the user specifications.

In addition to the detailed presentation of the underlying physics, thermodynamics,

mathematics and numerical approaches, this report provides a thorough discussion of the

various inputs that are common to all TOUGH+ numerical simulation of flow and

transport processes through porous media. These include the data needed for the

appropriate sizing of the dynamically dimensioning arrays of TOUGH+ v1.5,

identification of processes that may be included or omitted (e.g., diffusion, porosity-

permeability dependence, etc.), and descriptions of the discretized simulation domain in

terms of elements and connections, of the properties of the various porous media in the

domain, of the computational parameters that control the mathematical approaches and the

execution specifics, of the definition of the initial and boundary conditions, and of the

output options and specifications. As discussed earlier, data specific to the TOUGH+ v1.5

application options are not discussed here, as they are included in the corresponding

User’s Manuals along with sample problems.

6
 2.0 Concepts, Underlying Physics, and
Governing Equations

2.1. Modeled Processes and Underlying Assumptions

The TOUGH+ v1.5 general-purpose simulator can be used as the basis to model all the

known processes and phenomena associated with the flow and transport of fluids and heat

through porous and/or fractured media, such as:

(1) The flow of gases and liquids in the geologic system

(2) The corresponding heat flow and transport

(3) The partitioning of the mass components among the possible phases

(4) Heat exchanges due to

a. Conduction

b. Advection/convection

c. Radiation

d. Chemical reactions

7
 e. Latent heat related to phase changes (ice melting or water fusion, water

evaporation or vapor condensation)

f. Gas dissolution

(5) Equilibrium or kinetic chemical reactions,

(6) The multi-phase transport of solutes and colloids, accounting for advection,

molecular diffusion, and sorption

(7) The effects of solutes on the system behavior

Note that this list is not comprehensive, but only indicative of some of the most

obvious applications of the TOUGH+ family of codes. A significant effort has been

invested in the incorporation into the code of the most recent advances in physics

thermodynamics, mathematics and numerical analysis, and in keeping the simplifying

assumptions involved in the development of the underlying models of the code to a

minimum. Thus, the main assumption involved in TOUGH+ is that the laws governing the

flow of fluids (Darcian and non-Darcian) and heat are known and valid in the simulated

domain under the conditions of the study.

In the present section we present some fundamental principles and governing

equations that are universally applicable to (and present in) all problems of flow and

transport investigated by the TOUGH+ family of codes. More specific equations that are

applicable to the particular problems investigated by the individual members of the

TOUGH+ family (each describing a particular equation of state – EOS) will be presented

in detail and discussed in the User’s Manual accompanying each of these application

options.

8
2.2. The Mass and Energy Balance Equation
Following Pruess et al. [1999; 2012], mass and heat balance considerations in every

subdomain (gridblock) into which the simulation domain is been subdivided by the

integral finite difference method dictates that

d
∫ M κ dV = ∫ F κ ⋅ n dà + ∫ q κ dV , (2.1)
dt n
V Γn Vn

where:

V, Vn volume, volume of subdomain n [L3];

M κ
mass accumulation term of component κ [kg m-3];

A, Γn surface area, surface area of subdomain n [L2];

F κ
Darcy flux vector of component κ [kg m-2s-1];

n inward unit normal vector;

q κ
source/sink term of component κ [kg m-3s-1];

t time [T].

2.3. Mass Accumulation Terms

Under equilibrium conditions, the mass accumulation terms M for the mass components
κ

in equation (2.3) are given by

Mκ = ∑ φ Sβ ρ β X κ , κ ≡ 1,..., Nκ (2.2)
β
β ≡1,..., N β

where

φ porosity [dimensionless];

9
 3
ρ β density of phase β [kg m− ];

Sβ saturation of phase β [dimensionless];

X κβ mass fraction of component κ in phase β [kg/kg]

Nκ number of components κ in phase β [kg/kg]

Nβ number of phases β [kg/kg]

By convention, in this document phases are denoted by capital letters in TOUGH+

v1.5, while components are denoted by lower case letters. The phases that are involved in

the most common of the TOUGH+ applications will be Aqueous (A), Gaseous (G) and

liquid Organic (O). Other possible phases may be Ice (I), Liquid CO2, Hydrates (H), etc.

Common components include water (w), gaseous species (g) such as air, CH4, CO2, etc.,

oil (o) and/or solutes (s).

2.4. Heat Accumulation Terms

The heat accumulation term includes contributions from the rock matrix and all the

phases is given by the equation

M θ =( 1− φ)ρ RC RT + ∑ φ Sβ ρ βU β , (2.3)
β =1,..., N β

where

ρR rock density [kg m-3];

CR heat capacity of the dry rock [J kg-1 K-1];

Uβ specific internal energy of phase β [J kg-1];

10
 The specific internal energy of the gaseous phase is a very strong function of

composition, is related to the specific enthalpy of the gas phase HG, and is given by

# P&
UG = ∑ X Gκ uGκ +U dep %% = H G − (( , (2.4)
κ =1,..., Nκ $ ρG '

where uGκ [J/kg] is the specific internal energy of component κ under the conditions of the

gaseous phase, and Udep is the specific internal energy departure of the gas mixture [J/kg].
€ The internal energy of the aqueous phase accounts for the effects of gas and solute

solution, and is estimated from

U A = X Aw uAw + ∑ X Aκ (uκA + U sol

κ
), (2.5)
κ =1,...,Nκ
κ ≠w

where uAw and uκA [J kg-1] are the specific internal energies of H2O and of all other

components κ at the conditions prevailing in the aqueous phase, respectively, and U sol
κ

€
[J/kg] are the specific internal energies corresponding to the dissolution of components κ

(other than H2O). The term uκA is determined from

T
P P
uκA = hAκ − = ∫ C dT − ρ
κ (2.6)
€ ρκ T0 κ

where T0 is a reference temperature, hAκ is the specific enthalpy of component κ, and C isκ

€
the temperature-dependent heat capacities of component κ [J kg-1 K-1].
€
The development of the equations of mass and energy balance of any other phase

(e.g., liquid organic, liquid CO2, solid ice, etc.) is entirely analogous.

11
2.5. Flux Terms

The mass flux Fκ [kg/m2/s] of a component κ includes contributions from all mobile

phases present in the system, i.e.,

Fκ = ∑ Fβκ , where Fβκ = X βκ Fβ , κ ≡1,..., N κ , (2.7)

β =1,..., N mβ

and N mβ is the number of mobile phases. Fβ is the flux of phase [kg/m2/s], and can be

described by several equations, the most common of which is Darcy’s law:

k rβ ρ β
Fβ = −k
µβ
(∇P − ρ g) ,
β β (2.8)

where

k rock intrinsic permeability [m2];

k rβ relative permeability of the aqueous phase [dimensionless];

µβ viscosity of the aqueous phase [Pa s];

Pβ pressure of the aqueous phase [Pa];

g gravitational acceleration vector [m s-2].

Note that Darcy’s law is applicable to flows in which the Reynolds number NR < 1,

i.e., when the fluid flow is laminar, and when there are negligible flow slippage effects.

This covers the overwhelming majority of problems of flow and transport through

geologic porous media. Turbulent flows and slippage effects (occurring in low-

permeability media) require other equations, which are discussed in detail in the User’s

Manuals of the specific TOUGH+ options describing such problems.

12
 The Darcy’s law in Equation (2.8) indicates a linear relationship between flux and

pressure differential between two points in space. An important point is that Equation

(2.8) involves the phase pressures, which (and their interrelationships) need to be properly

defined. In general, the relationship between the phases of any two phases is described by

the general equation

Pwβ = Pnβ + Pcnw , (2.9)

where Pwβ is the pressure of the wetting phase, Pnβ is the pressure of the non-wetting

phase, and Pcnw is the capillary pressure between the two phases (by convention a negative

number in TOUGH+ v1.5). The gas phase is always a non-wetting phase, but the wetting

behavior of all other phases depends on the surface chemistry of the porous medium, i.e.,

its affinity for a given phase. Usually, unconsolidated media and reservoir rocks are

water-wet, i.e., the aqueous phase preferentially coats the grains surfaces, to which it is

attached very strongly as indicated by a very large contact angle. Some petroleum

reservoir rocks (especially older ones, with long contact of the rock grains with an organic

phase) are oil-wet, and there are also rocks of mixed wettability.

For example, in a two-phase (aqueous and gas) system, the aqueous pressure PA is

given by

PA = PG + PcGW , where PG = ∑ PGκ . (2.10)

κ =1,..., Nκ

Here, PG is the gas pressure [Pa], PcGW is the capillary pressure [Pa] between the two

phases, and PGκ are the partial pressures [Pa] of the various gaseous component κ in the

gas phase, respectively.

13
 The solubility of a gaseous component κ in the aqueous phase is related to PGκ

through Henry’s law,

PGκ = H κ X Aκ , (2.11)

where H κ = H κ ( T , mAs ) [Pa] is the temperature- and salt concentration-dependent Henry’s

factor, and mAs is the molality of the dissolved salts in the aqueous phase. TOUGH+

incudes a library of such H κ functions that describe the water solubility of the 11 gases in

its internal gas property database.

The performance of these Henry’s factors has been determined to be very

satisfactory over a wide spectrum of applications that cover an extended temperature and

salinity range. However, more demanding problems involving high pressures, gas

mixtures and the presence of more complex electrolytes cannot be adequately represented

by the simple H κ factors described above. In these cases, TOUGH+ provides an option

of a more accurate (but more computationally demanding) estimation of the gas solubility

in water from the equality of fugacities in the aqueous and the gas phase. Such an option

is available only in specific TOUGH+ options.

In TOUGH+, the solubility of various components (ionic and non-ionic, such as

electrolytes and organic substances, respectively) in the aqueous phase is determined by

appropriate species-specific equations of solubility as functions of temperature and

pressure. Similarly, the solubility of various components into non-aqueous phases (e.g.,

an organic liquid phase in petroleum reservoirs, or a liquid CO2 phase) is described by

species- and phase-specific equations. With the exception of the H κ -based solubility of

gases in the aqueous phase (from Equation (2.11), which is available as a standard in the

14
TOUGH+ options that require it), the solubilities of substances in phases are discussed in

detail in the User’s Manuals of the corresponding application options.

The mass flux of the gaseous phase ( β ≡ G ) incorporates advection and diffusion

contributions, and is given by

krG ρG κ
FGκ = −kG
µG
( )
X G ∇PG − ρG g + J κG , κ ≡1,..., N κ (2.12)

where

! b$
kG = k0 #1+ & ; (2.13)
" P%

kG medium permeability to gas [m2];

k0 absolute permeability at large gas pressures or in liquid flow [m2];

b Klinkenberg [1941] b-factor accounting for gas slippage effects [Pa];

krG relative permeability of the gaseous phase [dimensionless];

µG viscosity of the gaseous phase [Pa.s].

Equations (2.12) and (2.13) introduce a non-linearity in the flow equation, which is

no longer Darcian and can now account for gas slippage effects. However, this non-

linearity is easy to implement, is only important in low-permeability media, and is

available as an option in its standard implementation (which involves a constant

Klinkenberg parameter b) in the TOUGH+ core code. Estimates of b can be obtained

from the tables listed in Wu et al. [1998] and various equation options listed in Moridis

and Freeman [20014], e.g., the equation proposed by Jones [1972]:

−0.36
b !# k $
&
= , (2.14)
bref #" kref &
%

15
in which the subscript ref denotes the known properties of a reference medium. More

complex gas slippage scenarios involving other methods for computing a variable b are

available in specific TOUGH+ options, and they are discussed in detail in the

corresponding User’s Manuals (e.g., see Moridis and Freeman (2014]).

The term J κG is the diffusive mass flux of component κ in the gas phase [kg/m2/s],

and is described by

JκG =−φτ 0τ G DGκ ρG ∇XGκ , κ ≡1,..., N κ (2.15)

!
τ TG

where DGκ is the multicomponent molecular diffusion coefficient of component κ in the

gas phase in the absence of a porous medium [m2/s], τTG is the total gas tortuosity, τ0 is the

medium-related component of tortuosity and τG is the gas-saturation related component of

tortuosity (both dimensionless). TOUGH+ includes several methods to compute the τ0

and τ β of a mobile phase β. If a constant value τ0 is not provided as input, then the

1 7
default is the model of Millington and Quirk [1961], according to which τ 0 = φ 3 Sβ 3 . The

various methods to compute τ β are discussed in Section 5.1.

The diffusive mass fluxes of the various components κ in the gas phase are related

through the relationship of Bird et al. [1960]

∑ J κG = 0 , (2.16)
k=1,..., Nκ

which ensures that the total diffusive mass flux of the gas phase is zero with respect to the

mass average velocity when summed over the gaseous components. Then the total gas

phase mass flux is the product of the Darcy velocity and density of the gas phase.

16
 Similarly, the flux of a dissolved species (e.g., salt – denoted by the s superscript in

the following equation – dissolved in the aqueous phase) is described by

s s s
FAs = X As FA + J sA , where J A =−φ SA τ!τ DA ρ A ∇X A ,
0 A (2.17)
τ TA

DAs is the
€ molecular diffusion coefficient of the dissolved species s in water, and τA is the

tortuosity of the aqueous phase. Note that mechanical dispersion is not accounted for in

Equations (2.12) and (2.17), and is not discussed in this description of the core TOUGH+

v1.5 code, but the subject is fully addressed in the User’s Manuals of TOUGH+ options in

which dispersion is important.

The heat flux accounts for conduction, advection and radiative heat transfer, and is

given by

Fθ = −kθ ∇T + fσ σ 0 ∇T 4 + ∑ hβ Fβ , (2.18)
β ≡1,..., N mβ

where

kθ composite thermal conductivity of the rock-fluids ensemble [W m-1 K-1];

hβ specific enthalpy of phase β ≡1,..., N mβ [J kg-1];

fσ radiance emittance factor [dimensionless];

σ0 Stefan-Boltzmann constant [5.6687 × 10-8 J m-2 K-4].

Similar to Equation (2.4), the specific enthalpy of the gas phase is computed as

HG = ∑ X Gκ hGκ + H dep , (2.19)

κ =1,..., Nκ

17
where hGκ is the specific enthalpy of component κ in the gaseous phase, and Hdep is the

specific enthalpy departure of the gas mixture [J kg-1]. The specific enthalpy of the

aqueous phase is estimated from

H A = X Aw hAw + ∑
κ =1,..., Nκ
( κ
)
X Aκ hAκ + H sol , (2.20)
κ ≠w

where hAw and hAκ [J kg-1] are the specific enthalpies of H2O and of all other components κ

at the conditions prevailing in the aqueous phase, respectively, and H sol

κ
[J kg-1] are the

specific internal energies corresponding to the dissolution of components κ (other than

H2O). The development of the equation of the specific enthalpy of any other phase (e.g.,

liquid organic, liquid CO2, solid ice, etc.) is entirely analogous.

2.6. Source and Sink Terms

In sinks with specified mass production rate, the withdrawal of mass component κ is

described by

q̂κ = ∑ X βκ qβ , κ ≡1,..., N κ (2.21)

β ≡1,..., N mβ

where qβ is the mass production rate of the mobile phase β [kg/m3]. For a prescribed

production rate, the phase flow rates qβ are determined internally according to different

methods (e.g., the relative mobility at the element where the sink is located) available in

TOUGH+. For source terms (well injection), the addition of a mass component κ occurs

at desired (and known) rates q̂ κ ( κ ≡1,..., N κ ).

18
 The corresponding heat exchange Qθ associated with the addition or withdrawal

of mass at any given source or sink is described by

# &
Q = ∑
θ % ∑ X β qβ hβ (
κ κ(
(2.22)
%
κ ≡1,..., Nκ $ β ≡1,..., N mβ '

where hβκ is the specific enthalpy of the mass component κ that is partitioned in mobile

phase β [J/kg] – see Equations (2.19) and (2.20).

2.7. Thermophysical Properties

2.7.1. Water

The properties and parameters of liquid water and steam in TOUGH+ are provided by (a)

fast regression equations based on data from NIST [2000] and (b) steam table equations

from the most recent IAPWS97 formulations [Wagner et al., 2000; IAPWS, 2007]. These

equations are accurate up to 2000 oC and 100 MPa, and computationally more efficient

that those in the earlier versions of TOUGH+ (i.e., those in Moridis et al. [2008]). The

viscosity and the thermal conductivity of the water substance in its various phases were

computed using the recent correlations of thermal conductivity of IAPWS [2008] and

IAPWS [2011a], respectively.

The enthalpy, sublimation pressure and fusion/melting pressure of ice (on the ice-

vapor and ice-liquid water equilibrium lines of the water phase diagram), as well as the

densities of the ice and of the liquid water in these regions, are computed from the

correlations of IAPWS [2009; 2011b; 2012]. The thermal conductivity of ice was

19
computed using the heat capacity polynomial equation with the coefficients reported in

Yaws [1999].

2.7.2. Real Gases

The properties of the gas phase are provided by one of the three cubic equations of state

that are available in the supplemental code unit T_RealGas_Properties.f95 (see

Section 3.6) of TOUGH+ v1.5: (a) the Peng-Robinson equation [Peng and Robinson,

1976], (b) the Soave-Redlich-Kwong equation [Soave, 1972], and the standard Redlich-

Kwong equation [Redlich and Kwong, 1949]. This package includes a database of the

fundamental properties of 12 gases, and it computes their compressibility, density,

fugacity, specific enthalpy and internal energy (ideal and departure) over a very wide

range of pressure and temperature. Additionally, the package computes the gas viscosity

and thermal conductivity using the method of Chung et al. [1988], and binary diffusivities

from the method of Fuller et al. [1969] and Riazi and Whitson [1993]. TOUGH+ allows

computation of all these properties not only in pure gases (i.e., involving a single gaseous

component), but also in gas mixtures of either constant composition (in which the gas

phase can be treated as a single pseudo-component of fixed composition) or of variable

composition (in which case the various gaseous components are tracked individually).

This real-gas package in TOUGH+ also allows determination of gas solubility in

water either by using a set of temperature-dependent Henry’s coefficients, or by equating

fugacities in the gas and aqueous phases through a process that involves the computation

of the activity coefficients (in the aqueous phase) and of the fugacities. For most

applications involving low-solubility gases (especially single-component ones) or

20
relatively low P and T, extensive experience with scoping calculations has indicated that

the temperature-dependent Henry’s coefficient H m can provide reliable estimates of gas

solubility. Henry’s method is less reliable in the case of dissolution of multi-component

gases in the presence of electrolytes (such as salts) of high ionic strength. In these cases,

determination of gas solubility through the fugacities and activity coefficients provides the

necessary accuracy of the dissolution estimates, albeit at the cost of a significant larger

computational load. Only the solubility option that is appropriate for the problem at hand

is activated in the various TOUGH+ application options that involve coexistence of gas

and aqueous phases.

2.7.3. Other Phases and Components

The physical and thermal properties of all other phases and components that may be

involved in the various TOUGH+ applications cannot be generalized. Thus, they are

described by (a) appropriate computation methods that are coded in the corresponding

TOUGH+ application option, and (b) by the corresponding parameter values that are

provided in the TOUGH+ input file. These are described fully in the User’s Manuals of

the individual members of the TOUGH+ family involving such phases and components.

21
2.8. Porosity and Intrinsic Permeability
Changes
2.8.1. Changes caused by P- and T-dependence

The effect of pressure change on the porosity of the matrix is described by three options.

The first involves the standard equation

φ = φ0 FPT , FPT = exp [α P ΔP + αT ΔT ] ≈ 1+ α P ΔP + αT ΔT , (2.23)

where ΔP=P-P0, ΔT=T-T0, the subscript ‘0’ denotes a reference state, αT is the thermal

expansivity of the porous medium (1/K) and αp is the pore compressibility (1/Pa), which

can be either a fixed number or a function of pressure [Moridis et al., 2008; 2009; 2012].

The second option describes the P-dependence of φ as a polynomial function of P. The

third option (discussed in detail in Section 2.10) describes the φ-dependence on P in

unconsolidated media that gain significant mechanical strength from the presence of solid

phases such as ice or hydrates (see Moridis [2014]). The pore compressibility αP of such

media is a function of the saturation SS of the solid phase(s). The thermal dependence of

φ is still described by the exponential factor exp "#αT (T − T0 )$% .

The φ - k relationship in the matrix is described by the general expression of

Rutqvist and Tsang [2003] as:

k ( "φ %+
= exp *γ $ −1'- , (2.24)
k0 ) # φ0 &,

where γ is an empirical permeability reduction factor that ranges between 5 (for soft

unconsolidated media) and 29 (for lithified, highly consolidated media). Note that the

equations described here are rather simple and apply to matrix φ and k changes when the

22
changes in p and T are relatively small. These equations are not applicable when large

pressure and temperature changes occur in the matrix, cannot describe the creation of new

(secondary) fractures and cannot describe the initiation, propagation and the

characteristics of fractures as the fluid pressures, the temperatures, the fluid saturations

and the stresses change. For such problems, it is advisable to use the T+M model [Kim

and Moridis, 2013] that couples the flow and thermal processes in the various TOUGH+

options with the ROCMECH geomechanical code. This coupled model accounts for the

effect of changing fluid pressures, saturations, stresses, and temperatures on the

geomechanical regime and provides an accurate description of the evolution of φ and k

over the entire spectrum of P and T covered during the simulation.

2.8.1. Changes caused by the evolution of solid phases

When solid phases are deposited in a porous medium, through chemical precipitation or

freezing of pore fluids, the ability of the porous medium to transmit fluids can change

profoundly. The deposition of solids in a porous medium reduces the void space available

for fluids. Such reduction in porosity will give rise to a reduction in permeability as well.

There is an extensive literature, going back to the 1920s, about the manner in

which permeability declines as portions of the pore space are filled by solids, and a

bewildering variety of porosity-permeability correlations have been obtained from

experimental and theoretical studies [Scheidegger, 1974, and references therein; Morrow

et al., 1981; Vaughan, 1987; Verma and Pruess, 1988; Phillips, 1991; Pape et al., 1999;

Xu et al., 2004]. Within the scope of the work undertaken here it is not possible to

perform a thorough review of different permeability reduction models, and to evaluate

23
their suitability for representing permeability reduction due to formation of hydrate and/or

ice. Dearth of relevant information prevents considering whether any of the models

developed for solid precipitation in porous media are valid to systems involving solid

phases such as ice or hydrates. Instead, this brief discussion addresses salient features of

pore channels to highlight the most important effects, and then explain the rationale

behind the preliminary choices made in this study.

It is obvious that permeability effects depend not just on the magnitude of porosity

change, but on geometric properties of the pore channels, and on where and how solid

deposition in those channels occurs. The lack of unanimity among different investigators

about the correlation between porosity and permeability change reflects the great diversity

of pore channel geometries and precipitation processes in porous media. The simplest

models conceptualize porous media as bundles of capillary tubes, which gives rise to a

simple power law dependence of permeability k on porosity φ,

n
k !φ $
= Fφ S = # & (2.25)
k0 " φ0 %

where the term F S is a permeability adjustment factor that describes the effects of (a) the
φ

presence of solid phases other than the medium grains (such as ice, hydrate or

precipitating salts), and (b) changes in porosity on permeability, and the subscript “00”

denotes properties at a reference state.

The exponent n typically will be in the range from 2 to 3 [Phillips, 1991], giving a

rather mild dependence of permeability on porosity. However, in media with inter-

granular porosity, pore channels generally have a convergent-divergent geometry,

24
consisting of a succession of ‘pore throats’ with small radius and ‘pore bodies’ with large

radius (Figure 2.1).

If solids are deposited uniformly along the pore walls, or are preferentially

deposited in the throats, then even relatively minor amounts of deposition can give rise to

a dramatic decrease in permeability. Such behavior has been observed in field and

laboratory-scale systems, including the diagenesis of sandstones [Pape et al., 1999],

precipitation around geothermal injection wells [Xu et al., 2004], and hydrothermal flows

in laboratory specimen [Morrow et al., 1981; Vaughan, 1987].

In these systems, a rather modest amount of precipitate, that leaves most of the

original pore space available for fluids, nonetheless caused order-of-magnitude changes in

absolute permeability. Such behavior can be understood from ‘tubes-in-series’ models of

pore space, as shown in Figure 2.2 [Verma and Pruess, 1988].

Figure 2.1. Schematic of pore channels, showing convergent-divergent geometry with a

succession of pore throats and pore bodies.

25
Figure 2.2. Tubes-in-series model of pore channels.

If one assumes that solids are deposited as a layer of uniform thickness on the pore

walls, then permeability will be reduced to zero when the throats become clogged, while

plenty of (disconnected) porosity remains in the pore bodies. This leads to the concept of

a non-zero ‘critical porosity’ φc at which k is reduced to zero, with a permeability

reduction as given in Equation (2.25) [Verma and Pruess, 1988; Xu et al., 2004].

n
k " φ − φc %
= Fφ S = $ ' (2.26)
k0 # φ0 − φc &

Fractal models also give a very strong dependence of permeability on porosity,

with exponents in relationships such as Equation (2.25) as large as n = 10 or more [Pape

et al., 1999]. The above discussion clearly indicates the need fundamental research for the

determination of the sites within the porous media at which solid phases (such as ice and

hydrates) form preferentially.

26
2.9. Multiphase Diffusion

2.9.1. General Considerations

Here we expand on the subject of diffusion, which has already been referred to (and

included in a mathematical from) in Equations (2.12) to (2.17). The discussion here, as

well as the diffusion formulation and its treatment in TOUGH+ v1.5, hews closely to

those in TOUGH2 [Pruess et al., 1999; 2012].

Molecular diffusion plays a minor role in many subsurface flow processes, but

may become a significant and even dominant mechanism for mass transport when

adjective velocities are small and/or the time frame of the simulated process is long.

Diffusive flux is usually written as being proportional to the gradient in the concentration

of the diffusing component (Fick’s law)

J = − d ∇C (2.27)

where d is an effective diffusivity, which in general will depend on properties of the

diffusing component, the pore fluid, and the porous medium. The concentration variable C

may be chosen in a number of different ways, e.g., mass per unit volume, moles per unit

volume, mass or mole fraction, etc. [Bird et al., 1960; de Marsily, 1986], with an

appropriate selection of units for the corresponding diffusivity d.

The basic Fick’s law in Equation (2.27) works well for diffusion of tracer solutes

that are present at low concentrations in a single-phase aqueous solution at rest with

respect to the porous medium. However, many subtleties and complications arise when

multiple components diffuse in a multiphase flow system. Effective diffusivities in general

may depend on all concentration variables, leading to non-linear behavior especially when

27
some components are present in significant (non-tracer) concentrations. Additional

nonlinear effects arise from the dependence of tortuosity on phase saturations, and from

coupling between advective and diffusive transport. For gases, the Fickian model has

serious limitations even at low concentrations, which prompted the development of the

“dusty gas” model that entails a strong coupling between advective and diffusive transport

[Mason and Malinauskas, 1983; Webb, 1998], and accounts for molecular streaming

effects (Knudsen diffusion) that become very important when the mean free path of gas

molecules is comparable to pore sizes. Further complications arise for components that are

both soluble and volatile, in which case diffusion in aqueous and gaseous phases may be

strongly coupled via phase partitioning effects. An extreme case is the well-known

enhancement of vapor diffusion in partially saturated media, which is attributed to pore-

level phase change effects [Cass et al., 1984; Webb and Ho, 1998a, b].

2.9.2. Diffusion Formulation in TOUGH+

Because of the difficulties mentioned above, it is not possible to formulate a model for

multiphase diffusion that would be accurate under all circumstances. The pragmatic

approach used in Equations (2.15) and (2.17) describes the diffusive flux of component κ

in mobile phase β (= liquid, gas) by the general equation

Jκβ = −φτ 0τ β ρ β Dβκ ∇X βκ , (2.28)

where all terms are as previously defined. For ease and simplicity, it is convenient to

introduce a single diffusion strength factor that combines all material constants and

tortuosity factors into a single effective multiphase diffusion coefficient, as follows:

28
 D βκ = φτ 0τ β ρ β Dβκ (2.29)

For general two-phase conditions involving an aqueous and a gas phase, the total diffusive

flux is then given by

J κ = −D Aκ ∇X Aκ − D Gκ ∇X Gκ (2.30)

The saturation dependence of tortuosity is not well known at present. For soils the

Millington and Quirk [1961] model has frequently been used [Jury et al., 1983; Falta et

al., 1989], which yields non-zero tortuosity coefficients as long as the phase saturation is

non-zero. It stands to reason that diffusive flux should vanish when a phase becomes

discontinuous at low saturations, suggesting that saturation-dependent tortuosity should be

related to relative permeability; e.g. τβ(Sβ) ≈ krβ(Sβ).

For components that partition between liquid and gas phases, more complex

behavior may be expected. For example, consider the case of a volatile and water-soluble

compound diffusing under conditions of low gas saturation where the gas phase is

discontinuous. In this case we have krG(SG) = 0 (because SG < Srg), and krA(SA = 1- SG) <

1, so that a model equating saturation-dependent tortuosity to relative permeability would

predict weaker diffusion than in single-phase liquid conditions. For compounds with

significant volatility this would be unrealistic, as diffusion through isolated gas pockets

would tend to enhance overall diffusion relative to single-phase liquid conditions.

29
2.10. Wettability-Related Phenomena

2.10.1. Relative permeability

In multiphase flow, each fluid phase occupies only part of the pore space, and its effective

permeability is reduced due to interference with the other phase(s). This effect is

represented by means of permeability reduction factors or relative permeabilities,

customarily denoted by krA and krG for liquid (aqueous) and gas, respectively, such that

effective permeability kβ to phase β (= A, G) is given by

kβ = k krβ , where k = k0 FφS = k00 krr FφS , (2.31)

and krr is the k relative magnitude that relates the permeability k0 of a given medium to k00

of the reference medium at the same P and T. The factor FφS – see Equation (2.25) –

denotes the effect of the evolution of solid phases (e.g., through salt precipitation or ice

formation) and is equal to one if no solid phases are present. The term krr is introduced to

account for situations in which the reference medium is different from the one under

consideration, as is often the case when insufficient data are available and parameter

estimation is based on scaling – such as the one described by Equation (2.33) – using

known media as references. For a reference medium different from the one under

consideration, krr = k0/k00. It is obvious that krr = 1 when the same medium is used as

reference.

The relative permeabilities are functions of the phase saturations Sβ (fraction of pore

space occupied by phase β), krβ = krβ (Sβ ) , and are usually obtained by measurement on

laboratory specimen of porous media. Figure 2.3 gives examples of commonly used

liquid and gas relative permeabilities. In Section 6, the reader can find a detailed

30
description of the various relative permeability options that are available in TOUGH+

v1.5 for the description of multi-phase flow.

krG = 1 - krA

krG krA

Swr 1-Sgr

Figure 2.3. Liquid and gas relative permeabilities based on the van Genuchten model [Finsterle,
1999].

2.10.2. Capillary pressure

Surface tension effects between different phases give rise to ‘capillary pressures’, denoted

by Pcap. Most mineral surfaces are preferentially wetted by water and, under partially-

31
saturated conditions, the pressure of the wetting (aqueous) phase inside a porous medium

will be less than that of the non-wetting (gas) phase (see Equation 2.9). The pressure

difference between the two phases is termed capillary pressure Pcap, because it relates to

the phenomenon of water level rise in a capillary tube. As indicated earlier, by convention

Pcap has a negative value in TOUGH+ v1.5.

Issues relating to capillary pressure can be conveniently discussed with reference

to the pore size distribution of the porous medium. Figure 2.4 shows a schematic

probability density function (pdf) for pore sizes, which expresses the probability p(r) of

having pores with radius r. At a given capillary pressure Pcap, pores with radius r' ≤ r

may be water-filled, where the cutoff radius r is related to the capillary pressure by the

Young-Laplace equation (reference)

2σ
Pcap = cos(ω ) , (2.32)
r

where σ is the surface tension (energy per unit surface area) at the water-gas interface, and

ω is the contact angle, which usually is close to zero for preferentially water-wet minerals.

In introducing Equation (2.30) we stated that pores with r' ≤ r may be water filled,

but whether indeed all pores with r' ≤ r will be water-filled at a prevailing capillary

pressure Pcap given Equation (2.30) – and those with r' > r will all be gas-filled – depends

upon issues of pore accessibility that are not captured by the pore size distribution. The

issue of accessibility arises because pores of certain radius may be entirely surrounded by

larger pores, or by smaller pores, so that either water may not enter them during a wetting

process, or may not be removed from them during a draining process. Accessibility gives

rise to the well-known phenomenon of capillary hysteresis, where at a given magnitude of

32
capillary pressure, water saturation will generally be larger during a drainage process than

during a wetting process [de Marsily, 1986]. Although capillary hysteresis is a well-

established effect, it is seldom taken into account in modeling applications; this is partially

due to numerical difficulties associated with it, partially because on the dependence of the

corresponding relative permeability and capillary pressure on the pathway to the current

state (as opposed to just the phase saturations), and partially because information on

applicable parameters is rarely available.

Figure 2.4. Schematic of probability density function p(r) for pore size distribution.

33
2.10.3. Effect of solid phase deposition on capillary pressure

The discussion in this section follows closely that in Moridis et al. [2008; 2009; 2012].

The effects of solids deposition become considerably more complicated, and

involve more than just permeability change, when multiple fluid phase are present, such as

water and gas. The capillary pressure Pcap may be profoundly altered when solids are

deposited.

Formation of solid phases will alter the pore size distribution, generally reducing

pore sizes, and thereby giving rise to stronger capillary pressures (see Equation 2.25). In

order to estimate these changes, we require information on the original pore size

distribution of the medium, and on the manner in which the pore size distributions will be

altered during solids deposition. As no such information is presently available for the

medium studied here, we proceed in a more phenomenological manner, and relate changes

in capillary pressures to overall changes in porosity and permeability of the medium.

Examining a variety of unconsolidated media, Leverett [1941] determined a dependence

of capillary pressure on permeability and porosity, as follows.

k00 φ
Pcap (SA ) = ⋅ Pcap,00 (2.33)
k φ00

where Pcap,00 is the capillary pressure corresponding to a reference medium at the

reference conditions, at which the permeability and porosity of the porous medium are

k00 and φ00, respectively. Equation (2.33) is used in the present analysis, in the following

manner. We represent active solids (e.g., ice that may melt or hydrate that may dissociate,

as opposed to solid minerals which are inert) by means of a solid saturation, denoted by

34
SS = SH +SI, which measures the fraction of active pore space occupied by solids. The

fraction of pore space available to fluid phases is SA + SG, and we have the constraint

SA + SG = 1 − SS (2.34)

Let Pcap,00 denote the capillary pressure function applicable to a porous medium free of

solid saturation (SS = 0) with a reference porosity φ00. The total current porosity φ and the

active (or available) porosity φa available to fluids are then defined by the equation

φ
φ a = ( FPT φrr φ00 ) (SA + SG ) ⇒ a = FPT φrr (SA + SG ) = FPT φrr (1-SS ) , (2.35)
   φ00
φ

where the term FPT is a porosity adjustment factor that accounts for the effects of pressure

and temperature on porosity, and φrr is the φ relative magnitude, which relates the porosity

φ0 of a given medium to the porosity φ00 of the reference medium at the same reference P

and T. This term is introduced to account for situations in which the reference medium is

different from the one under consideration, as is often the case when insufficient data are

available and parameter estimation is based on scaling using known media as references.

For a reference medium different from the one under consideration, φrr = φ0/φ00. It is

obvious that φrr = 1 when the same medium is used as reference.

When changes in P and T are not large in geomechanically stable media, then FPT

can be estimated from the following equation:

φ φ
FPT = = = exp [α P ΔP + αT ΔT ] ≈ 1 + α P ΔP + αT ΔT , (2.36)
φrr φ00 φ0

where ΔP = P - P0, ΔT = T - T0, αP and αT are the pore compressibility and thermal

expansivity, respectively (see discussion in Section 6.2). For large ΔP and/or ΔT in

compressible or geomechanically unstable media, FPT can be estimated from a full

35
geomechanical model that relates the resulting changes in geomechanical stresses and

strains to changes in porosity.

The argument in the capillary pressure function Pcap,0 on the r.h.s of Equation

(2.33) is the aqueous saturation SA, referred to total fluid porosity. We measure liquid

saturation on a scale that refers to total active (fluid plus hydrate- and ice-filled) pore

space in the ice- and hydrate-free porous medium. In the medium with solid saturation SS,

SA corresponds to a scaled saturation

SA
SA* = (2.37)
SA + SG

relative to fluid-filled pore space, and this is the value to be used in the estimation of

Pcap,0 in the r.h.s of Equation (2.44). In the next session we will discuss the estimation of

kr in the presence of solid phases (i.e., ice and/or hydrate) for use in Equation (2.44).
φ

2.10.4. Effect of solid phase deposition on relative permeability

From Equation (2.31), the partitioning of effective permeability to a fluid phase β into a

porous medium- and solid-saturation dependent part (k0 F S) and a fluid saturation-
φ

dependent part krβ is a matter of convention and convenience. It leads to a conceptual

ambiguity in the representation of permeability reduction from solid deposition in

multiphase flow. Indeed, such permeability reduction may be attributed either to a change

in absolute or intrinsic permeability (as described by the product k0 F S), as is done for
φ

single-phase flow, or it may be attributed to a change in the fluid relative permeability krβ.

When hydrate and/or ice forms inside a partially water-saturated porous medium,

such formation clearly must start in the water-filled portion of the pore space, but may not

36
remain limited to the water-filled portion, as the solid crystals may grow and extrude into

primarily gas-filled pores. In the absence of specific pore-scale information about where

hydrate and/or ice will likely form, it is not possible to ascertain the applicability of

relationships such as Equation (2.25) to the permeability reduction associated with hydrate

formation and/or freezing. Even if applicable, appropriate parameters for the problem

under study here are lacking.

We proposed two alternative models to describe the wettability processes (relative

permeability and capillary pressure) in hydrate- and/or ice-bearing media [Moridis et al.,

2008; 2011; 2012; Moridis, 2014]. The first model, hereafter referred to as the “Original

Porous Medium” (OPM) model, is based on the treatment of (a) porosity as unaffected by

the emergence of hydrates and/or ice (although subject to change due to changes in P and

T), (b) of the intrinsic permeability of the porous media as unchanging during the

evolution of the solid phases, and (c) of the fluid flow as a relative permeability issue

controlled by the saturations of the various phases in the pores. The second family of

models, hereafter referred to as the “Evolving Porous Medium” (EPM) models, considers

the evolution of the solid phases (hydrate and ice) as tantamount to the creation of a new

porous medium with continuously changing porosity and intrinsic permeability, the pore

space of which is occupied only by the two fluid phases (aqueous and gas).

2.10.4.1. The OPM model. This simpler model represents permeability reduction as

relative permeability effects, and does not require any new parameters to be introduced.

More specifically, this model assumes that in the presence of solid phase(s), relative

permeability to each fluid phase is given by the same function krβ(Sβ) as in the absence of

37
solids. This means that aqueous phase relative permeability krA = krA(SA) depends only

on the aqueous saturation SA, and is the same, regardless of how the remaining fraction (1-

SA) of the pore space is divided between gas and solid phases. A similar comment applies

to gas relative permeability krG = krG(SG).

Setting aside for a moment the issue of (kr krS), permeability reduction for the fluid
φ

phases then occurs simply because, when SS increases, fluid phase saturations SA and SG

generally must decrease also, as dictated by the constraint in Equation (2.32). This

prescription is tantamount to asserting that liquid phase flow behaves as though solids

deposition occurs entirely in what would otherwise be gas-filled pore space, while gas

phase flow behaves as though solids deposition occurs entirely in what otherwise would

be liquid-filled pore spaces. It is obvious that solids deposition cannot simultaneously

occur only in liquid and only in gas-filled pore spaces, which points to a limitation of the

proposed permeability reduction model. We nonetheless feel that a model that introduces

no new and uncertain parameters is preferable to a model that does.

The permeability adjustment factor is computed from the following expression:

Fφ S = krφ krS , (2.38)

where kr is the permeability φ-factor that describes the effect of changes in φ on

φ

permeability, and krS is the permeability S-factor that relates reduction in the intrinsic

permeability to the presence of solid phases (such as ice, hydrates or precipitating salts).

In the OPM model, krS = 1 by definition, and the permeability φ-factor in Equation

(2.31) can be computed as

38
 &1 when the effect of φ changes on k is neglected
(
krφ = ' (2.39)
(
)exp "#γ ( FPT −1)$% when the effect of φ changes on k is accounted for,

where γ is an empirical parameter [Rutqvist and Tsang, 2002], and FPT is computed from

Equation (2.36).

In the OPM model, Pcap are estimated from Equation (2.33), in which:

• φ/φ00 = φrr FPT is computed from Equation (2.36)

• k00/k = 1/krr kr is computed from Equations (2.31) and (2.38), and

φ

• Pcap,00 is computed based on the scaled saturations S* of Equation (2.37)

Thus, the final expression for estimating the capillary pressure in the OPM model is:

φrr FPT
Pcap = ⋅ ⋅ Pcap,00 (S * ) (2.40)
krr krφ

Additional scaling can be introduced by using the active porosity φa and φa/φ00 from

Equation (2.35) – as opposed to φ/φ00 from Equation (2.36) – in the computation of

Equation (2.33).

2.10.4.2. The EPM models. In recognition of, and attempting to overcome, the

limitations of the OPM permeability reduction model, we have proposed two EPM models

and performed sensitivity studies using the absolute (intrinsic) permeability modifications

that will be discussed in this section. While the EPM models provide valuable insights, a

more consistent and defensible model – based on both theoretical analyses and laboratory

and field studies – for the effects of emerging solid phases on fluid permeabilities should

be developed in the future.

39
 EPM Model #1. With intrinsic permeability modification based on relative

permeabilities, what absolute permeability should be used in the Leverett scaling Equation

(2.33). An attractive possibility would be to set k = k0 (krA + krG). This, however, is not

acceptable because the sum of liquid and gas relative permeabilities depends not just on

the solid saturation (SS), but on SA and SG individually.

As a plausible alternative, we consider the permeability reduction when the fluid-

available pore space is either entirely liquid-filled or entirely gas-filled, which leads to

krS = krA ( SA = 1− SS ) or krS = krG ( SG = 1− SS ) .

Note that either expression depends only on solid saturation SS. As an estimate of the krS,

we then take the average of the two,

1"
krS = #krA ( SA = 1− SS ) + krG ( SG = 1− SS )$% (2.41)
2

Equation (2.38) provides a simple estimate of the permeability φ-factor. Then the phase

effective permeabilities are computed using Equation (2.31), in which:

• kr is computed based on the scaled saturations from Equation (2.37),

β

• F S is computed from Equation (2.38),

φ

• kr is computed from Equation (2.39), and

φ

• krS is computed from Equation (2.41).

The capillary pressure in the EPM #1 model is estimated using Equation (2.44), in

which the various terms are computed as follows:

• φa/φ00, computed from Equation (2.35), is used instead of φ/φ00,

• k00/k = 1/krr F S is computed from Equations (2.38) and (2.38),

φ

• kr is computed from Equation (2.39), and

φ

40
 • krS is computed from Equation (2.41).

Thus, the final expression for estimating the capillary pressure in the EPM #1 model is:

φrr FPT (1− SS )

Pcap = ⋅ ⋅ Pcap,00 (S * ) (2.42)
krr krφ krS

EPM Model #2. The only difference between this model and the EPM #1 model is

in the equation used to estimate the krS term, with all other equations applying unchanged.

In the EPM#2 model, the quantity F S = kr krS in Equation (2.38) is provided by Equation
φ φ

(2.26), leading to

n
" φ (1-SS ) − φc %
krS = $ 0 ' (2.43)
# φ0 − φc &

The term kr is obtained from Equation (2.39). Thus, the effective permeabilities in
φ

the EPM #2 model are computed from Equation (2.31). Similarly, the capillary pressure

in the EPM #2 model is estimated using Equation (2.42) and krS from Equation (2.43).

2.10.5. Pore compressibility of unconsolidated media in the presence of

cementing solid phases

While the pore compressibility αP in Equation (2.23) can be considered as a constant or

even as a function of pressure during fluid flow through consolidated (lithified) porous

media and/or in unconsolidated media, this approach is inadequate when cementing solid

phases (such as ice and/or hydrates) are present in the pores. This is because the presence

of these solid phases imparts stiffness and increases the geomechanical strength of the

solid phase-impregnated porous medium, the porosity φ of which reacts much slower to

41
variations in pressure P. Thus, to accurately represent the evolution of φ as a function of

P in these cases needs to account for the effect of the saturation SS of such solid phases.

The most appropriate method for accounting for the effect of SS on the porosity of

unconsolidated media is by solving the coupled flow-geomechanical problem, estimating

variations in P, T and phase saturation, and computing the corresponding changes in

stresses and strains. These are then used to compute changes in φ and k. Such coupling is

a possibility in the TOUGH+ v1.5 code, which allows the use of the commercial

geomechanical model FLAC3D [ITASCA, 2002] to evaluate the interaction between flow

and geomechanical properties. This model is automatically invoked if the corresponding

executable file FLAC3D.exe is present in the TOUGH+ v1.5 directory and appropriate

inputs are provided to the TOUGH+ code (see detailed discussion in Section 5).

If the FLAC3D model is not available or is not invoked (a frequent choice, given

the large execution times required for such fully coupled flow-geomechanical problems),

it is possibly to describe the effect of cementing solid phases SS on the porosity φ and the

intrinsic permeability k of unconsolidated media by employing an empirical model that

describes the media compressibility as:

α P = exp {ln α PL + (ln α PU − ln α PL )"#1− Bx (2.25, 2.25, SS* )$%} , (2.44)

where

SS − SS min + δ
SS* = , (2.45)
SS max − SS min + 2δ

αPL is the lower limit of the medium compressibility (corresponding to the full

stiffening/strengthening effect of the presence of cementing solid phases such as ice

and/or hydrates), αPU is the upper limit of the medium compressibility (corresponding to

42
the absence of cementing solid phases), Bx is the incomplete beta function, SSmin is the

largest SS saturation at which αP = αPU, SSmax is the lowest SS saturation at which αP = αPL,

and δ is a smoothing factor. Equation (2.44) is based on geomechanical and geophysical

data derived from laboratory and field observations, and results in the curve of Figure 2.5

that scans between the αPU and the αPL compressibility limits.

The relative porosity φ/φ0 is estimated from Equation (2.36), which applies

unchanged, but with the composite compressibility αP computed from Equation (2.44).

Figure 2.6 shows the relationship between the relative porosity φ/φ0 and the pressure drop

ΔP in an unconsolidated medium, and describes the cementing effect of solid phases on

the medium behavior.

Figure 2.5. Compressibility of an unconsolidated porous medium impregnated with cementing

-8 -1
solid phases (ice and/or hydrates). In this example, SSmin = 0.15, SSmax = 0.4, αPU = 10 Pa , αPL =
-9 -1
10 Pa and δ = 0.015 [Moridis et al., 2008; 2009; 2011].

43
Figure 2.6. Effect of the varying compressibility described in Figure 2.5 on the porosity of an
unconsolidated porous medium undergoing depressurization for various levels of saturation SS of
cementing solid phases [Moridis et al., 2008; 2009; 2011].

44
2.11. Description of Flow in Fractured Media
The discussion in this section hews very closely to that of Pruess et al. [1999; 2012].

Figure 2.7 illustrates the classical double-porosity concept for modeling flow in fractured-

porous media as developed by Warren and Root [1963]. Matrix blocks of low

permeability are embedded in a network of interconnected fractures. Global flow in the

reservoir occurs only through the fracture system, which is described as an effective

porous continuum. Rock matrix and fractures may exchange fluid (or heat) locally by

means of ‘interporosity flow’, which is driven by the difference in pressures (or

temperatures) between matrix and fractures. Warren and Root approximated the

interporosity flow as being quasi-steady, with rate of matrix-fracture interflow

proportional to the difference in (local) average pressures.

Figure 2.7. Idealized double porosity model of a fractured porous medium [Pruess, 1983].

45
 The quasi-steady approximation is applicable to isothermal single-phase flow of

fluids with small compressibility, where pressure diffusivities are large, so that pressure

changes in the fractures penetrate quickly all the way into the matrix blocks. However, for

multiphase flows, or coupled fluid and heat flows, the transient periods for interporosity

flow can be very long (tens of years). In order to accurately describe such flows it is

necessary to resolve the driving pressure, temperature, and mass fraction gradients at the

matrix/fracture interface. In the method of “multiple interacting continua” (MINC)

[Pruess and Narasimhan, 1982; 1985; Pruess, 1983), resolution of these gradients is

achieved by appropriate subgridding of the matrix blocks, as shown in Figure 2.8. The

MINC concept is based on the notion that changes in fluid pressures, temperatures, phase

compositions, etc., due to the presence of sinks and sources (production and injection

wells) will propagate rapidly through the fracture system, while invading the tight matrix

blocks only slowly. Therefore, changes in matrix conditions will (locally) be controlled by

the distance from the fractures. Fluid and heat flow from the fractures into the matrix

blocks, or from the matrix blocks into the fractures, can then be modeled by means of one-

dimensional strings of nested grid blocks, as shown in Figure 2.8.

In general it is not necessary to explicitly consider subgrids in all of the matrix

blocks separately. Within a certain reservoir subdomain (corresponding to a finite

difference grid block), all fractures will be lumped into continuum # 1, all matrix material

within a certain distance from the fractures will be lumped into continuum # 2, matrix

material at larger distance becomes continuum # 3, and so on. Quantitatively, the

subgridding is specified by means of a set of volume fractions VOL(j), j = 1, ..., J,

into which the primary porous medium grid blocks are partitioned.

46
 Fractures

Matrix Blocks

Figure 2.8. Subgridding in the method of "multiple interacting continua" (MINC) [Pruess, 1983].

The MINC-process in the MeshMaker.f95 (a companion code distributed

with TOUGH+, see Section 7) operates on the element and connection data of a porous

medium mesh to calculate, for given data on volume fractions, the volumes, interface

areas, and nodal distances for a secondary fractured medium mesh. The information on

fracturing (spacing, number of sets, shape of matrix blocks) required for this is provided

by a proximity function PROX(x) which expresses, for a given reservoir domain Vo, the

total fraction of matrix material within a distance x from the fractures. If only two

continua are specified (one for fractures, one for matrix), the MINC approach reduces to

the conventional double-porosity method. Full details are given in a separate report

[Pruess, 1983].

The MINC-method as implemented in the MeshMaker.f95 code can also

describe global matrix-matrix flow. Figure 2.9 shows the most general approach, often

47
referred to as dual permeability, in which global flow occurs in both fracture and matrix

continua. It is also possible to permit matrix-matrix flow only in the vertical direction.

For any given fractured reservoir flow problem, selection of the most appropriate gridding

scheme must be based on a careful consideration of the physical and geometric conditions

of flow. The MINC approach is not applicable to systems in which fracturing is so sparse

that the fractures cannot be approximated as a continuum. A thorough discussion on the

various approached for the treatment of fractured media in TOUGH2 [Pruess et al., 1999;

2012] and in TOUGH+ v1.5 (which closely follows the TOUGH2 approach) can be found

in Doughty et al. [1999].

Figure 2.9. Flow connections in the “dual permeability” model. Global flow occurs between
both fracture (F) and matrix (M) grid blocks. In addition there is F-M interporosity flow [Pruess et
al., 1999].

48
 3.0. Design and Implementation of
TOUGH+ v1.5

3.1. Primary Variables

The thermodynamic state and the distribution of the mass components among the possible

phases in a TOUGH+ v1.5 simulation are determined from the equation of state (EOS)

solved in the specific TOUGH+ application option. Following the standard approach

employed in the TOUGH2 [Pruess et al., 1999] family of codes, in TOUGH+ the system

is defined uniquely by a set of Nκ primary variables (where κ denotes the number of mass

and heat components under consideration) that completely specifies the thermodynamic

state of the system [Pruess et al., 1999; 2012]. Nκ is the sum of the number of the mass

components (TOUGH+ being a compositional simulator) augmented by one – the heat

balance equation (see Section 2.2), with the T being (usually) the corresponding primary

49
variable that is included in the computations even in isothermal problems because it would

not be possible to estimate the fluid thermophysical properties of fluids without it.

The primary variables are the minimum number of independent variables,

knowledge of which permits the simultaneous solution for all the state (=dependent,

secondary) variables associated with the problem, thus uniquely defining the system. For

example, knowledge of pressure and temperature allows the complete definition of the

thermophysical properties of water in its liquid and vapor phases, thus P and T are

appropriate primary variables to define the aqueous and the vapor phase. However, these

are inappropriate primary variables during the liquid-vapor phase coexistence because

they are no longer independent, as there is a well-defined and unique relationship between

P and T along the saturation line.

The importance of selecting appropriate primary variables in the simulations of

fluid and heat flow processes cannot be overemphasized, especially when phase changes

are involved. Inappropriate selection of primary variables can lead not only to slow and

inefficient computations, but also to complete failure of the simulation for lack of

convergence. Although the number Nκ of the primary variables is initially set at the

maximum expected in the course of the simulation and does not change during the

simulation, the thermodynamic quantities used as primary variables can change in the

process of simulation to allow for the seamless consideration of emerging or disappearing

phases and components. This is because the change of phases and the evolution and

disappearance of components almost involves sensitivity to different parameters, which

inevitably necessitates change of the primary variables. Switching primary variables as

the state of the fluids changes has been standard practice in TOUGH2 [Pruess et al., 1999;

2012], and continues to be so in TOUGH+ v1.5. Experience thus far has indicated that the

50
change in primary variables, coupled with the selection of the appropriate ones, is a very

robust method that is capable of solving even the most demanding problems of multi-

phase, multi-component flow and transport in porous/fractured media.

The primary variables used in the various TOUGH+ v1.5 application options vary

with the type of problem EOS being solved and cannot be generalized. Tables of the

specific primary variables used for each state (phase co-existence) of fluids are listed in

the User’s Manuals of the application options that solve the corresponding problem.

3.2. Space and Time Discretization

The continuum equations (2.3) are discretized in space using the integral finite difference

method (IFD) [Edwards, 1972; Narasimhan and Witherspoon, 1976]. Introducing

appropriate volume averages, we have

∫ M dV = Vn M n , (3.1)
Vn

where M is a volume-normalized extensive quantity, and Mn is the average value of M

over Vn. Surface integrals are approximated as a discrete sum of averages over surface

segments Anm:

∫F κ
•n dΓ = ∑A F ,
nm nm
(3.2)
Γn m

Here Fnm is the average value of the (inward) normal component of F over the surface

segment Anm between volume elements Vn and Vm. The discretization approach used in the

integral finite difference method and the definition of the geometric parameters are

illustrated in Figure 3.1.

51
 The discretized flux is expressed in terms of averages over parameters for elements Vn

and Vm. For the basic Darcy flux term in Equation (2.15), we have

"k ρ % "P −P %
Fβ ,nm = − knm $ rβ β ' $ β ,n β ,m − ρ β ,nm gnm ' , (3.33)
$# µ β '&nm # Dnm &

where the subscripts (nm) denote a suitable averaging at the interface between grid blocks

n and m (interpolation, harmonic weighting, upstream weighting). Dnm = Dn + Dm is the

distance between the nodal points n and m, and gnm is the component of gravitational

acceleration in the direction from m to n. Discretization of diffusive fluxes raises some

subtle issues, and is discussed separately in Section 3.5.

Substituting Equations. (3.l) and (3.2) into the governing Equation (2.3), a set of first-

order ordinary differential equations in time is obtained.

d M nκ 1
dt
=
Vn
∑A nm
κ
Fnm + qnκ
m (3.4)

Time is discretized as a first-order finite difference, and the flux and sink and source

terms on the right-hand side of Equation (3.4) are evaluated at the new time level, tk+1 = tk

+ Δt, to obtain the numerical stability needed for an efficient calculation of strongly

nonlinear problems (such as the ones involving multiphase flow and phase changes). This

treatment of flux terms is known as fully implicit, because the fluxes are expressed in

terms of the unknown thermodynamic parameters at time level tk+l, so that these

unknowns are only implicitly defined in the resulting equations [Peaceman, 1977].

The time discretization results in the following set of coupled non-linear, algebraic

equations

52
 Δt $ '
Rnκ ,k+1 = M nκ ,k+1 − M nκ ,k − & ∑ Anm Fnm + Vn qn ) = 0
κ ,k+1 κ ,k+1
(3.5)
Vn % m (

where we have introduced residuals Rnκ ,k+1 . For each volume element (grid block) Vn, there

are N equations, so that for a system discretized into NE grid blocks, Equation (3.5)
κ

represents a total of Nκ × N E coupled non-linear equations.

3.3. The Newton-Raphson Iteration

The unknowns of Equation (3.5) are the Nκ × N E independent primary variables {xi; i = 1,

..., Nκ × N E } which completely define the state of the flow system at time level tk+l. These

equations are solved by Newton/Raphson iteration, which is implemented as follows. We

introduce an iteration index p and expand the residuals Rnκ ,k+1 in Equation (3.5) at iteration

step (p + 1) in a Taylor series in terms of those at index p, i.e.,

Figure 3.1. Space discretization and geometry data in the integral finite difference method.

53
 ∂ Rnκ ,k+1
Rnκ ,k+1 ( xi,p+1 ) = Rnκ ,k+1 ( xi,p ) + ∑ ∂x (x i,p+1
− xi,p ) + ... = 0 (3.6)
i i p

Retaining only terms up to first order, we obtain a set of N x NE linear equations for the κ

increments ( xi, p+1 − xi, p ) :

∂ Rnκ ,k+1
− ∑ (x i, p+1 − xi, p ) + ... = Rnκ ,k+1 ( xi, p ) (3.7)
i ∂ xi p

All terms ∂Rn/∂xi in the Jacobian matrix are evaluated by numerical differentiation.

Equation (3.7) is solved by sparse direct matrix methods or iteratively by means of

preconditioned conjugate gradients [Moridis and Pruess, 1995; 1998; Pruess et al., 1999;

2012]. Iteration is continued until the residuals Rnκ ,k+1 are reduced below a preset

convergence tolerance according to:

Rn,κ ,k+1
p+1
κ ,k+1
≤ ε1 (3.8)
M n, p+1

The default (relative) convergence criterion is ε1 = 10-5 (TOUGH+ input

parameter rel_convergence_crit, see Section 10). When the accumulation terms

are smaller than ε2 (TOUGH+ input parameter abs_convergence_crit, default ε2 =

1, see Section 10), an absolute convergence criterion is imposed,

Rn,κ ,k+1
p+1 ≤ ε1ε 2 (3.9)

The number of iterations to convergence varies with the nonlinearity of the

problem. For well-behaved problems, convergence is usually attained in 3-4 iterations. If

convergence cannot be achieved within a certain number of iterations (default = 8, see

Section 10), the time step size Δt is reduced and a new iteration process is started.

54
3.4. Implications of the Space Discretization Approach
It is appropriate to add some comments about the space discretization technique in

TOUGH+. The entire geometric information of the space discretization in Equation (3.5)

is provided in the form of a list of grid block volumes Vn, interface areas Anm, nodal

distances Dnm and components gnm of gravitational acceleration along nodal lines. There

is no reference whatsoever to a global system of coordinates, or to the dimensionality of a

particular flow problem.

The discretized equations are in fact valid for arbitrary irregular discretizations in

one, two or three dimensions, and for porous as well as for fractured media. This

flexibility should be used with caution, however, because the accuracy of solutions

depends upon the accuracy with which the various interface parameters in equations such

as (3.3) can be expressed in terms of average conditions in grid blocks. A general

requirement is that there exists approximate thermodynamic equilibrium in (almost) all

grid blocks at (almost) all times [Pruess and Narasimhan, 1985]. For systems of regular

grid blocks referenced to global coordinates in cylindrical (r,z) and/or Cartesian (x,y,z)

systems, Equation (3.5) is identical to a conventional finite difference formulation

[Peaceman, 1977; Moridis and Pruess, 1992].

3.5. Space Discretization of Diffusive Fluxes

Space discretization of diffusive flux in multiphase conditions raises some subtle issues. A

finite difference formulation for total diffusive flux, Equation (3.10), may be written as

55
 ( X ) −( X )
κ
A m
κ
A n ( X ) −( X )
κ
G m
κ
G n
(J )
κ
nm
= (D )
κ
A nm
Dnm
(
−D )
κ
G nm
Dnm
(3.10)

This expression involves the as yet unknown diffusive strength coefficients ( ΣκA )nm and

( ΣκG )nm at the interface, which must be expressed in terms of the strength coefficients in

the participating grid blocks. Invoking conservation of diffusive flux across the interface

between two grid blocks leads in the usual way to the requirement of harmonic weighting

of the diffusive strength coefficients.

However, such weighting may in general not be applied separately to the diffusive

fluxes in gas and liquid phases, because these may be strongly coupled by phase

partitioning effects. This can be seen by considering the extreme case of diffusion of a

water-soluble and volatile compound from a grid block in single-phase gas conditions to

an adjacent grid block that is in single-phase liquid conditions. Harmonic weighting

applied separately to liquid and gas diffusive fluxes would result in either of them being

zero, because for each phase effective diffusivity is zero on one side of the interface. Thus

total diffusive flux would vanish in this case, which is unphysical. In reality, tracer would

diffuse through the gas phase to the gas-liquid interface, would establish a certain mass

fraction in the aqueous phase by dissolution, and would then proceed to diffuse away from

the interface through the aqueous phase. Similar arguments can be made in the less

extreme situation where liquid saturation changes from a large to a small value rather than

from 1 to 0, as may be the case in the capillary fringe, during infiltration events, or at

fracture-matrix interfaces in variably saturated media.

TOUGH+ features the fully coupled approach employed in TOUGH2 [Pruess et

al., 1999; 2012], in which the space-discretized version of Equation (3.10) of the total

56
multiphase diffusive flux Equation (2.49) is re-written in terms of an effective multiphase

diffusive strength coefficient and a single mass fraction gradient. Choosing the liquid mass

fraction for this we have

#
( X ) − ( X ) &( ( X ) − ( X )
κ κ κ κ

(J )
κ
nm (∑ ) + (∑ )
= −%
%
$
κ

A
nm
κ

G
nm
G m G n

( X ) − ( X ) (' D
κ
A m
κ
A n
A m

nm
A n
, (3.11)

where the gas phase mass fraction gradient has been absorbed into the effective diffusive

strength term (in braces). Flux conservation at the interface then leads to the requirement

of harmonic weighting for the full effective strength coefficient. In order to be able to

apply this scheme to the general case where not both phases may be present on both sides

of the interface, we always define both liquid and gas phase mass fractions in all grid

blocks, regardless of whether both phases are present. Mass fractions are assigned in such

a way as to be consistent with what would be present in an evolving second phase.

This procedure is applicable to all possible phase combinations, including the

extreme case where conditions at the interface change from single-phase gas to single-

phase liquid. Note that, if the diffusing tracer exists in just one of the two phases,

harmonic weighting of the strength coefficient in Equation (3.11) will reduce to harmonic

weighting of either ΣκA or ΣκG . The simpler scheme of separate harmonic weighting for

individual phase diffusive fluxes is retained as an option.

3.6. Code Units of the TOUGH+ v1.5 Code

TOUGH+ v1.5 is written in standard FORTRAN 95/2003. It has been designed for

maximum portability, and runs on any computational form (Unix and Linux workstations,

57
PC, Macintosh) for which such compilers are available. Running TOUGH+ involves

compilation and linking the code created by combining

(a) the units of the core code with

(b) the supplemental code units, corresponding to the specific application

option that the simulator aims to address, and with

(c) code the units that are part of the TOUGH+ code ensemble but are invoked

only to carry out computations needed by the application option.

The units of the core TOUGH+ are the following:

(1) T_Allocate_Memory.f95
Code unit that is responsible for the dynamic memory allocation (following input
describing the size of the problem) and dimensioning of most arrays needed by the
code, in addition to memory deallocation of unnecessary arrays.

(2) T_Utility_Functions.f95
Code unit that includes utility functions (including a variety of mathematical
functions, table interpolation routines, sorting algorithms, etc.).

(3) T_Media_Properties.f95
Code unit that describes the hydraulic and thermal behavior of the geologic
medium (porous or fractured), i.e., multiphase capillary pressure and relative
permeability, and interface permeability, mobility and thermal conductivity.

(4) T_H2O_Properties.f95
Code unit that includes (a) all the water-related constants (parameters), and (b)
procedures describing the water behavior and thermophysical properties/processes
in its entire thermodynamic phase diagram. Because water is almost universally
present in geologic media, this is the most commonly used supplemental TOUGH+
v1.5 code unit.

58
(5) T_Geomechanics.f95
Code unit that describes the geomechanically-induced changes on the flow
properties of the porous media. These include porosity φ changes caused by
pressure and/or temperature variations, intrinsic permeability k changes caused by
porosity changes, and scaling of capillary pressures Pcap to reflect changes in φ and
k. The φ and k changes are computed using either simplified of full geomechanical
models. When the simplified model is invoked, φ is a function of (a) P and the
pore compressibility αP and (b) of T and the pore thermal expansivity αT, while (c)
k changes are estimated using empirical relationships (see Section 8). Changes in
φ and k can also be computed by using a full geomechanical model, which can be
optionally coupled with TOUGH+.

(6) T_Main.f95
Main program that organizes the calling sequence of the high-level events in the
simulation process, and includes the writing of important general comments in the
standard output files, timing procedures, and handling of files needed by the code
and/or created during the code execution.

(7) T_Matrix_Solvers.f95
A linear algebra package that includes all the direct and iterative solvers available
in TOUGH+ (see Section 10).

(8) T_Executive.f95
The executive unit of TOUGH+. It includes the procedures that advance the time
in the simulation process, estimate the time-step size for optimum performance,
populate the matrix arrays and invoke the solvers of the Jacobian, invoke special
linear algebra for matrix pre-processing in cases of very demanding linear algebra
problems, compute mass and energy balances, compute rates in sources and sinks,
compute binary diffusion coefficients, write special output files, and conduct other
miscellaneous operations.

59
(9) T_Inputs.f95
This code unit includes the procedures involved in the reading of the general input
files needed for TOUGH+ simulations. It does not include any procedure reading
the data needed by the application option (see later discussion).

All the code units listed above are common to all TOUGH+ simulations.

TOUGH+ also includes supplemental code units that are part of the wider TOUGH+ code

ensemble and are available for specialized computations needed by the various application

options. These are the following:

(1) T_RealGas_Properties.f95
Code unit that includes (a) a complete database of the parameter values of 12 gases
that are needed to estimate all their properties (see below) needed for TOUGH+
simulations, and (b) procedures describing the equation of state (EOS) of real
gases (pure or mixtures) using any of the Peng-Robinson, Redlich-Kwong, or
Soave-Redlich-Kwong cubic EOS model. The procedures in this code unit
compute the following parameters and processes: compressibility, density,
fugacity, enthalpy (ideal and departure), internal energy (ideal and departure),
entropy (ideal and departure), thermal conductivity, viscosity, binary diffusion
coefficients, solubility in water, and heat of dissolution in water.

(2) T_Salinity_Effects.f95
Code unit that computes all necessary properties and parameters in application
options that involve salinity (e.g., brines). It estimates the salt solubility in H2O,
the halite density and enthalpy, the effect of salinity on the density, viscosity and
enthalpy of the aqueous phase, as well as on the vapor pressure of H2O.

(3) T_NonDarcian_Flow.f95
Code unit that computes all parameters and variables needed for the application of
non-Darcian flow through porous and fractured media by accounting for inertial
(turbulent) or viscous (slippage) effects. Thus, this unit reads all the non-Darcian

60
 flow inputs, and then uses them to compute all the parameters of the turbulent flow
options (Forcheimer [1901] or Barre and Conway [2007]), of slippage flow
(Klinkenberg flow [Klinkenberg, 1941], Knudsen diffusion [Freeman et al., 2011]
or the Dusty Gas Model [Mason and Malinauskas, 1983; Webb, 1998]).

Finally, to develop a fully functional code, the core and supplemental code units of

TOUGH+ are combined with the application option-specific code units. These have a

standard function, structure and nomenclature (which uses the application option name

OptionName as part of the unit name), and are described below:

(1) T_OptionName_Definitions.f95
Code unit providing default parameter values describing the basic attributes of the
application option/equation of state (i.e., option name, number of components,
number of phases, etc.). At the initiation of the simulation, the default parameter
values supplied by this unit are compared to those provided by the user in the
standard input file to ensure correct dimensioning in the dynamic memory
allocation process (see Section 5.1).

(2) T_OptionName_Properties.f95
Code unit that describes the thermophysical properties and processes of phases and
components other than those available in the core and supplemental TOUGH+
codes (i.e., the water and the real gas properties). For example, such a code unit
would describe the oil properties or CO2 properties in application options that
would require such computations.

(3) T_OptionName_Specifics.f95
Code unit that includes procedures specific to the application option, such as the
reading of EOS-specific inputs, the preparation of EOS-specific output files, the
computation of mass and volume balances and the estimation of EOS-specific
parameters. Generic procedures and operator extension – which override
(overload) the standard procedures used by TOUGH+ for the simulation of non-

61
 hydrate problems – are defined in this code unit, which does not include any
procedures describing the equation of state involved in this application option.

(4) T_OptionName_EOS.f95
Code unit that describes the equation of state of the application option, assigns
initial conditions, computes the thermophysical properties of the medium and of
the phases (i.e., all the secondary variables), and determines phase changes and the
state of the system. This code unit also includes the procedure that computes the
elements of the Jacobian matrix for the Newton-Raphson iteration.

For example, the code units of the HYDRATE v1.5 application option [Moridis,

2014] are: T_Hydrate_Definitions.f95, T_Hydrate_Properties.f95,

T_Hydrate_Specifics.f95 and T_Hydrate_EOS.f95.

Additionally, TOUGH+ v1.5 is distributed with the MESHMAKER V1.5 code

(also written in FORTRAN 95/2003), which used to be part of the main code in the

TOUGH2 simulators [Pruess et al., 1999; 2012], but is a separate entity in the TOUGH+

family of codes. MESHMAKER is used for the space discretization (gridding) of the

domain of the problem under study (see Section 7).

NOTE: In compiling TOUGH+ v1.5, it is important that the free-format source code

option be invoked for proper compilation of the FORTRAN 95/2003 code.

62
 4.0. Input Data Requirements
and Structure

In this section, we discuss in detail the general input requirements for the TOUGH+ v1.5

simulations. These are the inputs that are common to any simulation, regardless of the

application option that is being used. The input data for TOUGH+ v1.5 are a superset of

the input data used in conventional TOUGH2 and older TOUGH+ simulations in order to

ensure backward compatibility (a functional requirement for TOUGH+ v1.5). While most

of the general inputs are similar in type, input format, and parameter representation to

earlier versions of the code, new input data structures and advanced formats are used for

the inputs for new capabilities that are unique to TOUGH+ v1.5, as well as for the inputs

of the TOUGH+ application options. The introduction of advanced constructs and

formats for the entire input data set is in progress, and these will be made available in

future TOUGH+ releases.

63
4.1. Input Procedure
The input procedure in the current version of TOUGH+ remains similar in many aspects

to that of TOUGH2 [Pruess et al., 1999; 2012] and earlier versions of TOUGH+ [Moridis

et al., 2008; 2009; 2012]. Input data can be provided in a flexible manner by means of

one or several ASCII data files. Unless otherwise indicated, all TOUGH+ inputs are in

standard metric (SI) units, such as meters, seconds, kilograms, ˚C and in the corresponding

derived units, such as Newtons, Joules, Pascal (= N/m2 for pressure), etc.

In the TOUGH+ standard input file, data are organized in data blocks that are

defined by keywords. Quite often, only the first five characters of the keywords typed in

columns 1-5 (see Table 4.1) are read, because these are sufficient to recognize the data

block. While the contents of the various blocks are described in detail in Sections 5-12,

here we describe some important records/keywords, and provide some general comments

about their occurrence and arrangement in the input file.

The user is directed to the Appendix, where a sample input file is listed for his/her

reference in the study of the input process.

4.1.1. Data Block/Keyword TITLE

The first record of the input file in any TOUGH+ simulation is TITLE, which includes a

header of up to 132 characters. This record is necessary for any simulation to begin.

4.1.2. Keyword/Record ENDCY

A record with the ENDCY keyword typed in columns 1-5 closes the TOUGH+ input file

and instructs the code to initiate the simulation.

64
4.1.3. Keyword/Record ENDFI

The presence of the ENDFI keyword in columns 1-5 is an alternative (to ENDCY) ending

keyword in a TOUGH+ standard input file. Presence of ENDFI keyword causes the

simulation to be skipped after printing basic input information. This is a useful option

when the simulation is limited to an attempt to obtain some basic information of the

properties and conditions of the system in its initial state.

4.1.4. Structure of TOUGH+ Standard Input Files

Every TOUGH+ v1.5 input file must (a) begin with the record/keyword TITLE and (b)

end either with the record ENDCY, or, alternatively, with the record ENDFI (if no flow

simulation is to be carried out). Data records beyond ENDCY (or ENDFI) are ignored.

Some data blocks, such as those specifying reservoir domains (ROCKS or

MEDIA), volume elements (ELEME), connections (CONNE), and sinks/sources (GENER),

have a variable number of records, while others have a fixed number of records. Unless

otherwise indicated, a blank record indicates the end of the variable-length data blocks.

The data block MEMORY must follow the data block TITLE because it provides all

the necessary information for the dynamic memory allocation and array dimensioning in

TOUGH+ v1.5. The data blocks between the MEMORY and the ENDCY/ ENDFI keywords

can be provided in arbitrary order, except for the data block ELEME, which must precede

(if either is present) the blocks CONNE and EXT-INCON. The blocks ELEME and CONNE

must either be both provided through the standard input file, or must both be absent, in

which case alternative means for specifying geometry data are employed (see Section 7).

65
The data block GENER can be omitted if there are no sinks and sources in the problem. If

the keyword START or RANDOMN is present (see Section 4.1.2), the data block INCON

can be incomplete, with elements (grid blocks) in arbitrary order, or can be absent

altogether.

Elements for which no initial conditions are specified in INCON are then assigned

domain-specific initial conditions from (a) block INDOM (if present), or (b) from the data

block EXT-INCON (if present), or (c) from the ‘generic’ initial conditions given in block

PARAM, along with default porosities given in block ROCKS. If START or RANDOMN is

not present, INCON must contain information for all elements, in exactly the same order

as they are listed in block ELEME.

Between data blocks, the standard TOUGH+ input file may include an arbitrary

number of records that do not begin with any of the TOUGH+ keywords. This is useful

for inserting comments about problem specifications directly into the input file. TOUGH+

v1.5 gathers all these comments and prints the first 50 such records in the standard output

file.

Much of the data handling in TOUGH+ is accomplished by means of disk files,

which can be edited and modified using any text editor. The initialization of the arrays for

geometry, generation, and initial condition data is always made from the disk files MESH

(or MINC), GENER, and INCON. A user can either provide these files at execution time,

or they can be written from TOUGH+ v1.5 input data during the initialization phase of the

program, or they can be obtained from the standard TOUGH+ v1.5 simulation outputs at

the end of a simulation for use in a continuation run. .

66
 If the data blocks GENER and/or INCON are not provided in the standard input file,

and if no disk files GENER and/or INCON are present, defaults take effect (no generation;

initial conditions from block INDOM, or from block EXT-INDOM, or defaults from block

PARAM). To ensure that these defaults are used and that erroneous inputs are not

introduced, the disk files GENER and/or INCON from a previous run must be removed

from the execution environment/directory. A safe way to use default generation and

initial conditions is to specify dummy data blocks in the input file, consisting of one record

with GENER or INCON, followed by a blank record.

The format of data blocks ELEME, CONNE, GENER, and INCON is basically the

same (see Section 7) when these data are provided as disk files and when they are given as

part of the input file. However, specification of these data through the input file rather

than as disk files offers some added conveniences, which are useful when a new

simulation problem is initiated. For example, a sequence of identical items (volume

elements, connections, sinks or sources) can be specified through a single data record.

Additionally, indices that are used internally for cross-referencing elements, connections,

and sources are generated internally by TOUGH+ rather than having them provided by the

user. The INCON, GENER, and INCON disk files written by TOUGH+ can be merged into

an input file without changes, keeping the cross-referencing information.

67
 Table 4.1. TOUGH+ v1.5 input data blocks

Keyword (+) Sec. Function

TITLE (1st record) 4.1.1 Data record (single line) with simulation title
MEMORY (2nd record) 5.1 Dynamic memory allocation
OptionName (#) Parameters describing properties and behavior of the specific
application option
ROCKS or MEDIA 6.2 Hydrogeologic parameters for various reservoir domains
RPCAP or 6.3 Optional; parameters for relative permeability and capillary
WETTABILITY pressure functions
DIFFUSION 6.4 Optional; diffusivities of mass components
*ELEME 7.1 List of grid blocks (volume elements)
*CONNE 7.2 List of flow connections between grid blocks
INDOM 8.1 Optional; initial conditions for specific reservoir domains
*INCON 8.2 Optional; list of initial conditions for specific grid blocks
EXT-INCON 8.3 Optional; list of initial conditions for specific grid blocks
BOUNDARIES 8.6 Optional; provides time-variable conditions at specific
boundaries
*GENER 9.1 Optional; list of mass or heat sinks and sources
PARAM 10.1 Computational parameters; time stepping and convergence
parameters; program options
SOLVR 10.2 Optional; specifies parameters used by linear equation solvers.
TIMES 11.2 Optional; specification of times for generating printout
SUBDOMAINS 11.3 Optional; specifies grid subdomains for desired time series
data
INTERFACES 11.4 Optional; specifies grid interfaces for desired time series data
SS_GROUPS 11.5 Optional; specifies sink/source groups for desired time series
data
ENDCY (last record) 4.1.3 Record closes TOUGH+ input file and initiates simulation
ENDFI (last record) 4.1.4 Alternative for closing TOUGH+ input file which causes flow
simulation to be skipped.
#: This described in the individual User’s Manual of the TOUGH+ v1.5 application option
*: Data can be provided as separate disk files and omitted from input file.
+: The bold face part of the keyword (left column) suffices for data block recognition

68
 Data describing time-variable boundary conditions can be entered in a tabular form

in the data block BOUNDARIES. In addition to the output files that are produced by any

TOUGH+ v1.5 simulation, TOUGH+ v1.5 provides the option of output files with time-

series data on variables that describe (a) the status of selected subdomains of the simulated

domain and (b) flows through user-defined interfaces within the domain and/or through

groups of sources and sinks (wells). The input data needed for the definition of these

subdomains, interfaces and well groups are provided through the data blocks

SUBDOMAINS, INTERFACES and SS_GROUPS, respectively (see Table 4.1). For

continuation runs, the presence of the output files corresponding to the SUBDOMAINS,

INTERFACES and SS_GROUPS data blocks – created during the last simulation that is to

be continued – is of critical importance because they contain information needed by the

TOUGH+ code for the seamless extension of the simulation period.

69
 PAGE LEFT INTENTIONALLY BLANK

70
 5.0. Memory Specification
and Allocation

5.1. Data Block MEMORY

This is the data block that reads the data for the dynamic memory allocation in the

TOUGH+ simulation. This block must always follow the TITLE record.

Record MEMORY.1

This record must always include the following header:

MEMORY

Record MEMORY.2

Reads the character variable EOS_Name according to FORMAT(A15). This

variable describes the Equation Of State (EOS) that describes the application
option simulated by TOUGH+ in the problem at hand. This parameter is used
strictly to allocate memory to the pertinent storage arrays and to compare the user-
specified values of the parameters provided in the next record (Record
MEMORY.3) against defaults specified in the code.

71
 Record MEMORY.3

The following information is provided in MEMORY.3 using a free format:

NumCom, NumEq, NumPhases, binary_diffusion

These parameters are defined as follows:

NumCom Integer denoting the number of mass components

(see Section 2)

NumEq Integer denoting the number of equations

NumPhases Integer denoting the number of phases.

binary_diffusion Logical variable indicating whether binary diffusion

is active (binary_diffusion=.TRUE.) or
ignored (binary_diffusion=.FALSE.).

The permissible combinations of the values of these parameters are discussed in

details in the User Manuals of each of the TOUGH+ application options.

Record MEMORY.4

In MEMORY.4, the following parameters are read using a free format:

coordinate_system, Max_NumElem,
Max_NumConx, ElemNameLength,
active_conx_only, boundaries_in_matrix

These parameters are defined as follows:

coordinate_system
Character variable describing the coordinate system used in the study. It
can assume the values 'Cartesian' or 'Cylindrical'.

Max_NumElem
Integer variable defining the maximum number of elements (cells,
gridblocks) in the discretized simulation domain

Max_NumConx
Integer variable defining the maximum number of connections in the
discretized simulation domain.

72
 Information on the concepts of elements and connections can be found in
Section 8, and in Pruess et al. [1999]. As a general rule, MaxNum_Conx
> ND*MaxNum_Elem, where ND is the dimensionality of the problem.

ElemNameLength
Integer variable defining the number of characters in the element names. It
may be either 5 or 8. The default in TOUGH+ v1.5.

active_conx_only
Logical variable indicating whether the simulation will be halted after
determining the active connections in the grid. This feature is useful when
running a simulation that uses a subset of the elements of a large grid
without correspondingly adjusting the connections, and is designed to
reduce the very large memory requirements for the connection-related
dynamic arrays.

When active_conx_only =.TRUE., the simulation stops once the

active connections (involving only the elements defined in the element list)
are determined. The active connections are stored in a new file called
Active_Connection_File. Then the simulation can be run using the
new connection list, thus having much lower memory requirements. For a
thorough discussion of elements and connections in TOUGH+, see Section
7.

boundaries_in_matrix
Logical variable indicating how inactive elements (describing constant-
conditions boundaries, see Section 8.5, and denoted either by a negative or
very large volume, or by setting the variable elem_activity='I ', see
Section 7.2) are to be treated when solving the equations of mass and
energy balance.

When boundaries_in_matrix =.TRUE., then all inactive elements

are assigned very large volumes (1050 m3) to maintain constant conditions
during the simulation, and are included in the Jacobian matrix. Otherwise,
only the active elements are included in the Jacobian.

This feature is useful when using older TOUGH2 input files, and when a
parallel version of the code is employed. With newly developed input files,
the boundaries_in_matrix =.FALSE. value is highly
recommended.

Record MEMORY.5

The integer variable Max_NumSS (declaring the maximum number of expected

sources and sinks) is read using a free format.

73
 Record MEMORY.6

In this record, the integer variable Max_NumMedia is read using a free format.
This variable represents the maximum number of geologic media with different
properties to be considered in the simulations.

Record MEMORY.7

In MEMORY.7, the following logical variables are read using a free format:

element_by_element_properties,
porosity_perm_dependence,
scaled_capillary_pressure,
Option_tortuosity_CompuMethod

These parameters are defined as follows:

element_by_element_properties
Logical variable (flag) indicating whether each gridblock has its own
hydraulic properties (φ and k), in which case they are read on an element-
by-element basis (see Sections 7 and 8). This feature is necessary (a) in the
simulation of very heterogeneous systems, and/or (b) when k changes in
response to pressure and/or temperature variations, or in response to
changes in the geomechanical regime of the system.

When element_by_element_properties =.FALSE., then the φ

and k of a particular element are determined from the general properties of
the corresponding porous medium (see Section 6). When
element_by_element_properties =.TRUE., then element-
specific φ and k are read as part of the initial conditions in the INCON data
block (see Section 8).

porosity_perm_dependence
Logical variable (flag) indicating whether the intrinsic permeability k in a
given element is to change as a function of changing porosity φ. As
discussed earlier, φ can change in response to P and/or T changes according
to relationships that are determined from simple or complex geomechanical
models (see Section 2).

When porosity_perm_dependence =.FALSE., then k is

unaffected by changes in φ. When porosity_perm_dependence
=.TRUE., then k is readjusted internally to reflect the effect of changes in
φ that are estimated using either an empirical model (see Equation 2.51,
Section 2) or a full geomechanical model.

74
 Note that when porosity_perm_dependence =.TRUE., the
variable element_by_element_properties is set internally to
.TRUE. because activation of the porosity_perm_dependence
feature results in element-specific hydraulic properties.

scaled_capillary_pressure
Logical variable (flag) indicating whether the capillary pressure Pcap will be
scaled to reflect variations in φ and k.

Activation of this feature by setting scaled_capillary_pressure

=.TRUE. may be needed (a) in highly heterogeneous systems in which
the element-specific properties vary significantly from those described by
the average (expected) values of the porous medium (as specified in
Section 6), and/or (b) when the significant variations in φ and k are
experienced in the course of the simulation (e.g., when
porosity_perm_dependence =.TRUE.).

Option_tortuosity_CompuMethod
A character variable describing the method of estimation of the binary gas
diffusivities. This input variable is important if one or more of the
following processes are considered in the simulation: turbulent flow,
diffusion, solute or colloid transport, a Dusty Gas Model [Mason and
Malinauskas, 1983; Webb, 1998]; otherwise it is ignored. The following
options are available:

='Relative_Perm': For domains for which a tortuosity parameter τ0 =

mediaTortu ≠ 0 is specified in data block MEDIA or ROCKS (see
( )
Section 6.1), τ β =τ β Sβ =krβ – see Equations (2.15) and (2.17).

='Saturation': For domains for which a tortuosity parameter τ0 =

mediaTortu ≠ 0 is specified in data block MEDIA or ROCKS (see
( )
Section 6.1), τ β =τ β Sβ =Sβ – see Equations (2.15) and (2.17).

='Constant': When this option is invoked, the constant τ0 =

mediaTortu values provided in the data block MEDIA or ROCKS
(Section 6.1) is used with no phase-saturation adjustment, i.e., τ β =1 .

NOTE: If mediaTortu=0 in data block MEDIA or ROCKS, then the

tortuosity τ0 is computed from the Millington-Quirk model in Equation
(2.21) as τ T β = τ 0τ β = φ 1/ 3Sβ10/ 3 . This value is then used in the computation
of the effective diffusion coefficients.

75
 Record MEMORY.8

In MEMORY.8, the following logical variables are read using a free format:

coupled_geochemistry, property_update

These parameters are defined as follows:

coupled_geochemistry
Logical variable (flag) indicating whether the simulation involves coupled
flow, thermal and geochemical processes. This feature is activated by
setting coupled_geochemistry =.TRUE. when geochemical
processes are considered in a TOUGH+ simulation involving coupled flow,
thermal and geochemical processes. The default value is
coupled_geochemistry =.FALSE..

property_update
Character variable (flag) indicating the type of property update when
coupled geochemical processes are involved in TOUGH+ simulations.
This variable is the same as in the next record (MEMORY.9), where it is
discussed in detail.

Record MEMORY.9

In MEMORY.9, the following logical variables are read using a free format:

coupled_geomechanics,
geomechanical_code_name,
property_update, num_geomech_param

These parameters are defined as follows:

coupled_geomechanics
Logical variable (flag) indicating whether the simulation involves coupled
flow, thermal and geomechanical processes. Activation of this feature by
setting coupled_geomechanics =.TRUE. indicates the use of a
complex geomechanical model to describe the relationship between
hydraulic media properties (φ and k) and geomechanical parameters (such
as stresses and strains). This complex geomechanical model overrides the
simplified models (based on pore compressibility and expansivity) that are
standard in TOUGH+ (see Section 2).

When this flag is invoked, the geomechanical model invoked by the

TOUGH+ code is that defined by the variable
geomechanical_code_name (see discussion of the next parameter).

76
 When coupled_geomechanics =.FALSE. (the default value), the
simplified geomechanical models are invoked, even if the executable of a
geomechanical code (named geomechanical_code_name) is present
in the TOUGH+ directory.

When coupled_geomechanics =.TRUE., and the executable of a

geomechanical code named geomechanical_code_name is present in
the TOUGH+ directory, the following files are created: (a) To_GMech,
containing the data (pressure, temperatures and phase saturations) that are
provided to the FLAC3D geomechanical code [Itasca, 2002] from the
TOUGH+ simulator for use in the computation of the geomechanical
properties of the system, and (b) the Fr_Gmech file, containing the data
(mainly stresses and strains) supplied by the FLAC3D code for use by the
TOUGH+ simulator for the computation of the variable (geomechanically-
dependent) hydraulic properties φ and k. These two files are necessary for
communication between the two codes, because the lack of shared memory
(disallowed because of intellectual property concerns) makes data
exchange by means of these two external files as the only viable option.
Additionally, the new file Init_Stress that stores the initial stresses
and strains is created if this is a new run; in a continuation run, the initial
(i.e., at t = 0, not at the time of the initiation of the continuation run)
stresses and strains are read from the old file In_Stresses.

geomechanical_code_name
A character variable of maximum length 6 that provides the name of the
executable of the geomechanical code that is coupled with the TOUGH+
code when coupled_geomechanics =.TRUE.. In the current
implementation of the code, the geomechanical model used in conjunction
with the common TOUGH+ application options is the FLAC3D
commercial simulator (ITASCA, 2002), i.e.,
geomechanical_code_name = FLAC3D. However, the TOUGH+
can easily accommodate any other geomechanical simulator that conforms
to its data exchange formats and requirements.

When coupled_geomechanics =.TRUE. and the executable of a

geomechanical code named geomechanical_code_name is present in
the TOUGH+ directory, it is treated as a C subroutine that is called by the
TOUGH+ simulator. If (a) coupled_geomechanics =.TRUE. and
(b) the geomechanical_code_name is different from FLAC3D , or
does not correspond to any executable available in the TOUGH+ directory,
or is blank, the TOUGH+ code resets coupled_geomechanics to
=.FALSE..

property_update
Character variable (flag) denoting the manner of property update as a result
of interdependent changes in the hydraulic (flow) and geomechanical

77
 properties. The variable property_update can assume the following
values: 'Continuous', indicating continuous property update (i.e., in
every Newtonian iteration of every timestep) and participation in the
Jacobian matrix; 'Iteration', indicating property updates in every
Newtonian iteration of every timestep, but without any contribution to the
Jacobian; and 'Timestep', indicating a single property update at the end
of each timestep and no contribution to the Jacobian. The option
property_update = 'Continuous' yields the most accurate
solutions that are accurate over any pressure range and media
geomechanical property range but results in longer execution times, while
the option property_update = 'Timestep' leads to faster solutions,
but which are acceptably accurate in less compressible media and for a
mild ΔP.

If a property_update value other than the three described above is

read, then an error message is printed and the simulation is aborted.

NumGeomechParam
Integer variable defining the number of geomechanical parameters that are
to be provided by the geomechanical code (to be read by the TOUGH+
simulator from the Fr_Gmech file), and which will be used to estimate
the updated φ and k. Geomechanical codes such as FLAC3D can provide a
very large number of geomechanical properties and parameters of interest,
but in the current version of the TOUGH+ only two (stress and strain) are
needed, and these are stored separetely. Thus, NumGeomechParam is not
needed and can be set to zero (indicating no additional memory utilization)
when the FLAC3D geomechanical model is invoked for hydrate
simulations that involve coupled geomechanical effects. However, the
parameter is available for generality and maximum flexibility in future
code developments that may include the use of additional geomechanical
simulators and constitutive equations.

5.2. Internal Checks

The data provided in Record MEMORY.3 (NumCom, NumElem and NumPhases)

define the dynamic array dimensioning of the arrays that describe all the primary and

secondary variables in TOUGH+ v1.5 simulations. These are compared to the default

value ranges of the same parameters for the specific TOUGH+ application option (EOS)

that are provided in the code unit T_OptionName_Definitions.f95 (see Section

78
3.6). If any of these parameter values fall outside the acceptable range, this is a grave

(unrecoverable) error; a detailed error message is printed into the standard output file and

the simulation ceases.

Similarly, an error message is printed and the simulation is aborted if the default

application option (EOS) name name(s) in T_OptionName_Definitions.f95

heading of the OptionName data block (see Table 4.1) that provides the application

option (EOS) name is in conflict with the.

79
 PAGE LEFT INTENTIONALLY BLANK

80
 6.0. Physical Properties of System

This section lists and describes the physical properties of the system that are common to

any TOUGH+ simulation, including the rock properties (block ROCKS or MEDIA), the

relative permeability and capillary pressure properties (block RPCAP), and the properties

specifying multi-component diffusion (block DIFFU). Additionally, it describes the

procedure for introducing block-by-block permeability modification.

6.1. Data Block ROCKS or MEDIA

This block introduces material parameters (flow and thermal) for all the different geologic

media (porous or fractured) involved in the domain simulated by the appropriate TOUGH

v1.5 application.

Record ROCKS.1

Format (A5, I5, 8E10.4)

81
 mediaName, NAD, mediaDensG,
mediaPoros, (mediaPerm(i), i = 1,3),
mediaKThrW, mediaSpcHt, PoMedRGrain

mediaName
Material name (rock type).

NAD
Integer variable; if zero or negative, defaults will take effect for a number
of parameters (see below);

≥ 1: will read another data record to override defaults.

≥ 2 and < 5: will read two more records with domain-specific parameters
for relative permeability and capillary pressure functions.

= 5: In addition to the four records read for NAD > 2, an additional record
will be read with the coefficients of the porosity polynomial φ/φ0 = F0 +
F1ΔP + F2ΔP2 +…+ FnΔPn, where φ0 is the reference (initial) default
porosity and ΔP = P-P0 is the deviation from the initial pressure P0. This
equation will be used instead of Equation (2.47) to estimate the effect of
pressure on the medium porosity.

= 6: In addition to the four records read for NAD > 2, an additional record
will be read with the coefficients of Equations (2.56) and (2.57) that
describe the compressibility of an unconsolidated porous medium in the
presence of cementing solid phases (such as ice and/or hydrates).

= 8: In addition to the four records read for NAD > 2, an additional record
will be read with information on the specifics of inertial or slippage effects
on fluid flow through this medium.

mediaDensG
Rock grain density [kg/m3]

mediaPoros
Default porosity φ0 (void fraction) for all elements belonging to domain
MediumName for which no other porosity has been specified in block
INCON. Option "START" is necessary for using the default porosity.

mediaPerm(i),i=1,…,3
Absolute permeabilities along the three principal axes, as specified by
ConxKi in block CONNE.

82
mediaKThrW
Formation heat conductivity under fully liquid-saturated conditions
[W/m/˚C].

mediaSpcHt
Rock grain specific heat [J/kg/˚C].

PoMedRGrain
Rock grain radius [m]. This is needed for the estimation of the surface
reaction area in some TOUGH+ v1.5 application options involving kinetic
chemical reactions. If PoMedRGrain = 0.0e0 (e.g., when no value is
provided), the TOUGH+ v1.5 code provides a grain radius estimate using
the Kozeny-Carman approximation. This variable is needed in very
specific TOUGH+ v1.5 applications options, and is discussed in the
corresponding User’s Manuals.

Record ROCKS.1.1 (optional, when NAD ≥ 1 only)

Format (11E10.4)

mediaCompr, mediaExpan, mediaKThrD,

mediaTortu, mediaKlink, mediaOrgCF,
mediaCritSat, mediaPermExpon, mediaBeta,
mediaGama, PhiZeroStress

mediaCompr
Pore compressibility α P = (1 φ ) (∂φ ∂ P )T [Pa-1], see Equation (2.47) –
default αP = 0.0.

mediaExpan
Pore expansivity αT = (1 φ ) (∂φ ∂ T ) P [1/ ˚C], see Equation (2.47) – default
αT = 0.0.

mediaKThrD Formation heat conductivity under desaturated

conditions [W m-1K-1] – default is mediaKThrD = mediaKThrW.

mediaTortu
Tortuosity factor for binary diffusion. If mediaTortu = 0, a porosity and
saturation-dependent tortuosity will be calculated internally from the
Millington and Quirk [1961] model, Equation (2.53).

mediaKlink
Klinkenberg parameter b [Pa-1] for enhancing gas phase permeability
according to the relationship of Eaution (2.13) – default is 0.

83
 mediaOrgCF
Not used.

mediaCritSat
Critical total mobile phase saturation at which the permeability of the
medium that experience precipitation of solids becomes equal to zero; it is
equal to the critical “open” porosity φc of a porous medium at which its
permeability becomes zero – needed only when the EPM model is invoked
(see Sections 2.8 and 2.10).

mediaPermExpon
Permeability reduction exponent for solid phase-bearing systems – needed
only when the EPM model is invoked, see see Sections 2.8 and 2.10,
Equation (2.43).

mediaBeta
The parameter β used for the computation of porosity as a function of
geomechanical stresses σ according to the equation:

φ = φ0 - φσ=0 {exp[β(σ - P)] - exp[β(σ - P0)]}

0

where φσ=0 is the medium porosity at zero stress, and the subscript 0
indicates initial conditions.

NOTE: This φ computational option can be invoked only when all of the
following conditions are met: (a) the option coupled_geomechanics
=.TRUE. is activated, i.e., when the TOUGH+ simulations are coupled
with a full geomechanical model such as the FLAC3D code, (b) at least one
medium β ≠ 0, and (c) all media pore compressibilities αP = 0.

If coupled_geomechanics =.FALSE., this parameter is ignored and

the variations in φ are computed from one of the other computational
options available in TOUGH+. The same occurs when
coupled_geomechanics =.TRUE., all media β = 0, and at the pore
compressibility of at least one medium αP ≠ 0.
mediaGama
The parameter γ used for the computation of intrinsic permeability k as an
empirical function of variations in the porosity φ – See Equation (2.24).

PhiZeroStress
The porosity at zero stress φ . This variable is used only when the option
σ=0

coupled_geomechanics =.TRUE. is activated, i.e., when the

TOUGH+ simulations are coupled with a full geomechanical model such as
the FLAC3D code. If coupled_geomechanics =.FALSE., this
variable is ignored.

84
Record ROCKS.1.2 (optional, NAD ≥ 2 only)

Format (I5, 5X, 7E10.4)

RelPermEquationNum, (RelPermParam(i),i=
1,…,7)

RelPermEquationNum
Integer parameter indicating the type of the relative permeability function
of the medium under consideration (see detailed discussion in Section
6.3.1).

RelPermParam(i),i=1,…,7
Real parameters corresponding to the relative permeability function
described by the RelPermEquationNum option (see Section 6.3.1).

Record ROCKS.1.3 (optional, NAD ≥ 2 only)

Format (I5, 5X, 7E10.4)

PcapEquationNum, (PcapParam(i),i=1,…7)

PcapEquationNum
Integer parameter indicating the type of the capillary pressure function of
the medium under consideration (see detailed discussion in Section 6.3.2).

PcapParam(i),i=1,…7
Real parameters corresponding to the capillary pressure function described
by the PcapEquationNum option (see Section 6.3.2).

Record ROCKS.1.4 (optional, NAD = 5 only, to be used when the media

porosity is described as a polynomial function of the pressure
change ΔP)

Format (I5, 5X, 7E20.13)

PhiPolyOrder, (PhiCoeff (i),i=0,…,6)

PhiPolyOrder
Order n of the polynomial φ/φ0 = F0 + F1ΔP + F2ΔP2 +…+ FnΔPn. For a
constant φ, PhiPolyOrder = 0.

PhiCoeff(i),i=0,…,6
Coefficients Fn (n = 0,…,PhiPolyOrder) of the φ = φ(ΔP) polynomial.

85
 Record ROCKS.1.4 (optional, NAD = 6 only, to be used when cementing solid
phases such as ice and/or hydrates are present in the pores of
unconsolidated media – see Section 2.11.3)

Format (10E10.4)

LoComp, SatAtLoComp,
HiComp, SatAtHiComp, DeltaSat

LoComp
The lower limit of the medium compressibility αPL [Pa-1],
corresponding to the full stiffening/strengthening effect of the
presence of cementing solid phases such as ice and/or hydrates –
see Equation (2.56).

SatAtLoComp
= SSmax, i.e., the lowest SS saturation at which αP = αPL – see Equations
(2.44) and (2.45).

HiComp
The upper limit of the medium compressibility αPU [Pa-1], corresponding to
the absence of cementing solid phases– see Equation (2.44).

SatAtHiComp
= SSmin, i.e., the largest SS saturation at which αP = αPU – see Equations
(2.44) and (2.5).

DeltaSat
The smoothing factor δ – see Equation (2.45). A value of δ = 0.015 is
suggested – see Figure 2.5.

Repeat records ROCKS.1, ROCKS.1.1, ROCKS.1.2, ROCKS.1.3 and

ROCKS.1.4 for all the porous/fractured media in the domain under investigation.

Record ROCKS.2
A blank record closes the ROCKS data block.

Note: The number of media described in the data block ROCKS/MEDIA cannot
exceed the number Max_NumMedia specified in the MEMORY data block (See
Section 5.1). If this happens, an error message is printed and the simulation is
aborted.

86
6.3. Data Block RPCAP

This block introduces information on relative permeability and capillary pressure

functions, which will be applied for all flow domains for which no data were specified in

records ROCKS.1.2 and ROCKS.1.3. A catalog of relative permeability and capillary

pressure functions is presented in Sections 6.3.1 and 6.3.2, respectively.

Record RPCAP.1

Format (I5, 5X, 7E10.4)

DefaultRelPermType, (RPD(i),i=1,7)

DefaultRelPermType
Integer parameter describing the type of the default relative permeability
function (see Section 6.3.1).

DefaultRelPermType, (RPD(i),i=1,7)
Real parameters corresponding to the relative permeability function
selected by the DefaultRelPermType parameter (see Section 6.3.1).

Record RPCAP.2

Format (I5, 5X, 7E10.4)

DefaultCapPresType, (CPD(I),i=1,7)

DefaultCapPresType
Integer parameter describing the type of the default capillary pressure
function (see Section 6.3.2).

(CPD(i), i=1,…,7)
Real parameters corresponding to the capillary pressure function selected
by the DefaultCapPresType parameter (see Section 6.3.2).

All the relative permeability and capillary pressure functions that are available in

TOUGH+ for use in the various application options are listed in the following sections.

87
6.3.1. Two-Phase Relative Permeability Functions

6.3.1.1. RelPermEquationNum = 1: Linear functions

krA increases linearly from 0 to 1 in the range RP(1) ≤ SA ≤ RP(3);
krG increases linearly from 0 to 1 in the range RP(2) ≤ SG ≤ RP(4)
Restrictions: RP(3) > RP(1); RP(4) > RP(2).

6.3.1.2. RelPermEquationNum = 2: Power functions

n
krA = ( SA ) , krG = 1 , where n = RP(1)

6.3.1.3. RelPermEquationNum = 3: Corey's curves [Corey,1954]

2
krA = Ŝ 4 , (
krG = 1 − Ŝ ) (1 − Ŝ )
2

where Ŝ =
( SA − SirA )
(1 − SirA − SirG )
with SirA = RP(1); SirG = RP(2)
Restrictions: RP(1) + RP(2) < 1

6.3.1.4. RelPermEquationNum = 4: Grant's curves [Grant, 1977]

krA = Ŝ 4 , krG = 1 − krA

where Ŝ =
( SA − SirA )
(1 − SirA − SirG )
with SirA = RP(1); SirG = RP(2)
Restrictions: RP(1) + RP(2) < 1

6.3.1.5. RelPermEquationNum = 5: All phases perfectly mobile

krG = krA = 1 for all saturations; no parameters

6.3.1.6. RelPermEquationNum = 6: Functions of Fatt and Klikoff (1959)

3 3
krA = ( S * ) , krG = (1 − S * )

88
 where: S* =
( SA − SirA )
(1 − SirA )
with SirA = RP(1).
Restriction: RP(1) < 1.

6.3.1.7. RelPermEquationNum = 7, -7: van Genuchten-Mualem model

[Mualem, 1976; van Genuchten, 1980]

# * λ 2

krA = $ {
% S 1 − (1 − [S * ]1 λ ) } if SA < SmxA
%&1 if SA ≥ SmxA

Gas relative permeability can be chosen from among several options. For
RelPermEquationNum = 7, it is computed from one of the following
three forms, of which the second is the Corey [1954] equation and the third
is the modified Stone equation [Stone, 1970] – see Section 6.1.3.9:

#
% 1 − krG if SirG =0
% 2
krG (
= $ 1 − Ŝ
%
) (1 − Ŝ ) 2
if SirG >0 and nG ≤ 0
nG
%
& (S ) if SirG >0 and nG > 0

For RelPermEquationNum =-7, the gas relative permeability is

computed from the model of Parker et al. [1987]:

( ( λ +2
* " * 1/ λ %
/ 1 − S )$1− ( S ) ' , if 1− S * >0
krG = ) *# & -
/
* 0 if 1− S * ≤ 0

The krA and krG estimates are subject to the following restrictions:

0 ≤ krA ,krG ≤ 1

SA − SirA S −S
Here, S* = and S! = G irG
SmxA − SirA 1−SirA

89
 Parameters: RP(1) = λ
RP(2) = SirA
RP(3) = SmxA
RP(4) = SirG
RP(5) = nG

Notation: Parameter λ is m in van Genuchten’s notation, with m = 1 - 1/n.

6.3.1.8. RelPermEquationNum = 8: Function of Verma et al. [1985]

krA = Ŝ 3 , krG = A + B Ŝ + C Ŝ 2

where Ŝ =
( SA − SirA )
( SmxA − SirA )
Parameters as measured by Verma et al. [1985] for steam-water flow in an
unconsolidated sand:
SirA = RP(1) = 0.2
SmxA = RP(2) = 0.895
A = RP(3) = 1.259
B = RP(4) = -1.7615
C = RP(5) = 0.5089

6.3.1.9. RelPermEquationNum = 9: Modified version of Stone’s first three-

phase relative permeability method [Stone, 1970].

(* (*" n ,,
SA − SirA % **
krA = max )0, min )$ ' ,1-- ,
*+ +*# 1− SirA & .**.
(* (*" nG ,,
SG − SirG % **
krG = max )0, min )$ ' -- ,
,1
*+ *+# 1− SirA & *.*.
krH = 0

Parameters are SirA = RP(1), SirG = RP(2), n = RP(3), nG = RP(4).

Note: When RP(4)=0.0e0, it is reset internally to RP(4)=RP(3) .

90
 If the user has access to the source code, he/she may modify the source code of

TOUGH+ (segment T_Media_Properties.f95) to include other relative

permeability relationships. However, for this task the user needs to be familiar with the

concepts of generic processes (and overloading) in object oriented programming

languages.

6.3.2. Two-Phase Capillary Pressure Functions

6.3.2.1. PcapEquationNum = 1: Linear function

#
% − CP(1) for SA ≤ CP(2)
%
%%
Pcap = $ 0 for SA ≤ CP(2)
%
% CP(3)− SA
% − CP(1) for CP(2)< SA <CP(3)
%& CP(3)−CP(2)

Restriction: CP(3) > CP(2)

6.3.2.2. PcapEquationNum = 2: Function of Pickens et al. [1979]

1x
( "A %+
Pcap
* #B
(
= − P0 ) ln $ 1+ 1− B 2 A 2 ',
&- )
with
! S $ ( S − SirA ) SA
A = #1+ A & A0 , B = 1−
" SA0 % ( SA0 + SirA ) SA0

where P0 = CP(1), SirA = CP(2), SA0 = CP(3), x = CP(4)

Restrictions: 0 < CP(2) < 1 ≤ CP(3); CP(4) ≠ 0

6.3.2.3. PcapEquationNum = 3: TRUST capillary pressure [Narasimhan et

al., 1978]

91
 ( 1η
* − P − P " 1− SA % for SA <1
e 0$ '
** # SA − SirA &
Pcap = )
*
* 0 for SA <1
*+

where P0 = CP(1), SirA = CP(2), η = CP(3), Pe = CP(4)

Restrictions: CP(2) ≥ 0; CP(3) ≠ 0

6.3.2.4. PcapEquationNum = 4: Milly’s function [Milly, 1982]

1/ 4
A
" 0.371 %
Pcap =−97.783×10 , A = 2.26 $$ −1''
# SA − SirA &

where SirA = CP(1)

Restriction: CP(1) ≥ 0

6.3.2.5. PcapEquationNum = 6: Leverett’s function [Leverett, 1941; Udell

and Fitch, 1985]

Pcap = − P0 ⋅ σ (T )⋅ f (SA )

where

σ(T): surface tension of water (supplied internally in TOUGH+)

2 3
f (SA ) = 1.417 (1− S * ) − 2.120 (1− S * ) +1.263 (1− S * )

S* =
( SA − SirA )
(1 − SirA )
Parameters: P0 = CP(1), SirA = CP(2)

Restriction: 0 ≤ CP(2) < 1

92
6.3.2.6. PcapEquationNum = 7: van Genuchten function [van Genuchten,
1980]

Pcap = − P0 "$( S * ) −1%'

−1/ λ 1−λ
, S* =
( SA − SirA )
# & ( SmxA − SirA )
subject to the restriction: −Pmax ≤ Pcap ≤ 0

Parameters: CP(1) = λ = 1 - 1/n

CP(2) = SirA (should be chosen smaller than the
corresponding parameter in the relative
permeability function; see note below.)
CP(3) = 1/P0 = α/ρwg
CP(4) = Pmax
CP(5) = SmxA

Notation: Parameter λ is m in van Genuchten’s notation, with m = 1 - 1/n.

Note on parameter choices: In the van Genuchten’s derivation [1980] of the

capillary pressure equation, the parameter SirA for irreducible water
saturation is the same in the relative permeability and capillary pressure
functions. As a consequence, for SA → SirA we have krA → 0 and Pcap →−∞ ,
which is unphysical because it implies that the radii of capillary menisci go
to zero as liquid phase is becoming immobile (discontinuous). In reality, no
special capillary pressure effects are expected when liquid phase becomes
discontinuous. Accordingly, we recommend to always define a smaller SirA
for the capillary pressure as compared to that for the relative permeability
function.

6.3.2.7. PcapEquationNum = 8: Brooks-Corey [1964] equation modified to

account for effect of solid phases on capillary pressure

Pcap = − F ⋅ G ⋅ PGE (S * )v , S * =
( SA − SirA ) , F = 1+ A ⋅ Bx(a, b, SH )
(1 − SirA )
where
v = exponent with the following restrictions: v< 0 and |v| ≤ 1;
PGE = gas entry pressure;

93
 G = error function equation that smoothes curve near S=0;
F = factor that describes effect of hydrate on capillary pressure;
A = parameter > 0;
Bx = incomplete beta function;
a,b = input arguments for Bx;
SS = sum of solid saturations.

Parameters: CP(1) = PGE, CP(2) = v, CP(3) = SirA,

CP(4) = Pcap,max, CP(5) = A, CP(6) = a, CP(7) = b

6.3.2.8. PcapEquationNum = 9: No capillary pressure

Pcap ≡ 0 for all saturations; no parameters.

6.4. Data Block DIFFUSION

This block reads multicomponent diffusion coefficients using a NAMELIST format. This

is a very powerful format that allows maximum clarity and flexibility, accepting free

formats, arbitrary ordering of variables, insertions of comments anywhere in the input

fields, and providing the option of ignoring any of the NAMELIST parameters by not

assigning a value to it. For more information, the reader is directed to a textbook on

FORTRAN 95/2003.

Record DIFFUSION.1

This record includes general data describing key diffusion parameters. The
namelist in this record is named Diffusion_Key_Parameters, and has the
following general form.

&Diffusion_Key_Parameters
gas_diffusivity_equation_exponent = x.xEx,
P_at_RefDiffusivity = x.xEx,
Tk_at_RefDiffusivity = x.xEx
full_multiphase_diffusion = .x.
/

The parameters in the namelist Diffusion_Key_Parameters are defined as

follows:

94
gas_diffusivity_equation_exponent
A double precision variable describing the dependence of gas diffusivity on
temperature (see Equation 6.4). The default value is 1.80.

P_at_RefDiffusivity
Pressure at the reference diffusivity (in Pa). If P_at_RefDiffusivity
<= 0, the default value is 105 Pa.

Tk_at_RefDiffusivity
Temperature at the reference diffusivity (in K). If
T_at_RefDiffusivity <= 0, the default value is 273.15 K.

Option_gas_diffusivity_CompuMethod
A character variable describing the method of estimation of the binary gas
diffusivities. The following options are available:
='Standard': This option involves the application of Equation (6.4),
and requires non-zero multicomponent gas diffusivity values read from the
standard input file.
='Real_Gas_EOS': In this case, the binary gas diffusivities are
computed from the cubic equation of state used to determine all the real gas
properties. The diffusivities in the aqueous phase still need to be provided.
='Constant': When this option is invoked, the constant
multicomponent diffusivity values provided in the input file are used.

full_multiphase_diffusion
A logical variable describing the method of estimation of the method of
estimation of multiphase diffusive fluxes. The following options are
available:
=.TRUE.: With this option, harmonic weighting to the full multiphase
effective diffusion strength is applied. This includes contributions from gas
and aqueous phases, accounts for coupling of diffusion with phase
partitioning effects, and can describe the most general cases of diffusion
across phase boundaries.
=.FALSE.: In this case, harmonic weighting is performed separately for
the diffusive fluxes in the mobile phases.

Records DIFFUSION.2.1, DIFFUSION.2.2, etc.

Record DIFFUSION.2.1 is followed by DIFFUSION.2.x records, with x =

1,…,NubMobPhases (i.e., the number of mobile phases in the system under
study). These records describe component diffusivities in the various phases. The
same namelist is used in each one of these records. It is named

95
 Component_Diffusivities_in_Phases, and has the following general
form:

&Component_Diffusivities_in_Phases
phase = x,
phase_number = x,
component(1) = x,
component_number(1) = x,
component_diffusivity(1) = x.xEx,
component(2) = x,
component_number(2) = x,
component_diffusivity(2) = x.xEx,
…
…
…
/

The parameters in the namelist Diffusion_Key_Parameters are defined as

follows:

phase
A character variable identifying the mobile phase for which the
diffusivities of the various components are reported. The possible options
in the T+H code are 'Aqueous' and 'Gas'.

phase_number
An integer variable providing the number of the phase in the phase
numbering sequence used in the code. For example, the possible values in a
TOUGH+ v1.5 option involving aqueous and gas phases are:
= 2 for phase = 'Aqueous', and
= 1 for phase = 'Gas'.

component
A character array of dimension NumCom (see Section 5.1) identifying the
various mass components partioned in the phase in question (denoted by
phase). The possible options the example above could be 'CH4', 'H2O' and
'NaCl' (if salinity is considered).

component_number
An integer array providing the number of the component in the numbering
sequence used in the code. The possible options in the T+H code are:
= 1 for component = 'CH4'
= 2 for component = 'H2O'
= 3 for component = 'NaCl' (if present)

96
 component_diffusivity
A double precision array of dimension NumCom (see Section 5.1)
describing the value of the multicomponent diffusivities Dβκ (see Equations
(2.59) and (6.4)) of the various components κ in the phase β under
consideration (indentified by phase and phase_number, respectively).

NOTE: The records DIFFUSION.2.x must provide € data for all mobile
phases and all components, even if the gas diffusivities may be overridden
internally when Option_gas_diffusivity_CompuMethod =
'Real_Gas_EOS'.

The structure of the namelists Diffusion_Key_Parameters and

Component_Diffusivities_in_Phases (and their use as input formats in
the data block DIFFUSION) are best illustrated in the example of Figure 4.1.

DIFFUSION-----*----2----*----3----*----4----*----5----*----6----*----7----*----8
&Diffusion_Key_Parameters gas_diffusivity_equation_exponent = 1.8d0
P_at_RefDiffusivity = 1.0d5, ! in Pa
Tk_at_RefDiffusivity = 273.15d0, ! in K
Option_gas_diffusivity_CompuMethod = 'Real_Gas_EOS',
full_multiphase_diffusion = .TRUE.
/
&Component_Diffusivities_in_Phases
phase = 'Aqueous', phase_number = 2,
component(1) = 'CH4', component_number(1) = 1,
component_diffusivity(1) = 1.0d-10, ! (m2/s) ! Diffusivity of component 1 in phase 2
component(2) = 'H2O', component_number(2) = 2,
component_diffusivity(2) = 1.0d-10, ! (m2/s) ! Diffusivity of component 2 in phase 2
component(3) = 'NaCl', component_number(3) = 3,
component_diffusivity(3) = 1.0d-10 ! (m2/s) ! Diffusivity of component 3 in phase 2
/
&Component_Diffusivities_in_Phases
phase = 'Gas', phase_number = 1,
component(1) = 'CH4', component_number(1) = 1,
component_diffusivity(1) = 1.0d-05, ! (m2/s) ! Diffusivity of component 1 in phase 1
component(2) = 'H2O', component_number(2) = 2,
component_diffusivity(2) = 1.0d-05, ! (m2/s) ! Diffusivity of component 2 in phase 1
component(3) = 'NaCl', component_number(3) = 3,
component_diffusivity(3) = 0.0d-00 ! (m2/s) ! Diffusivity of component 3 in phase 1
/

Figure 4.1. The DIFFUSION data block, with examples of the Diffusion_Key_Parameters
and Component_Diffusivities_in_Phases namelists

97
6.4.1. User Options for Multiphase Diffusion

The treatment of full multiphase diffusion requires parameter specifications in two

different data blocks. First, the user must set binary_diffusion to a value of

.TRUE. in data block MEMORY. Diffusivities are input through the data block DIFFU

(Section 6.4). See note in Section 6.4 regarding an alternative approach, which allows

user to only specify the diffusion coefficient of salt and thus save in computation costs.

For two gaseous components (NG = 2), the multicomponent diffusivities in the gas

phase are the binary gas diffusivities. For NG > 2 gaseous components κi, the

multicomponent diffusivities of gas component κi are computed from the binary gas

diffusivities by the Wilke method [API, 1977] as

κi 1−YGκ i
DG
= NG
κn , (6.4)
YGκ n
∑ κiκn
n=1,n≠i dG

where dGκ iκ n is the binary gas diffusivity of gasesous components κ n and κ i .

When computing binary gas diffusivities dGκ iκ n using the option

Option_gas_diffusivity_CompuMethod = 'Real_Gas_EOS', their pressure

and temperature dependence is automatically accounted for in the computations. When

Option_gas_diffusivity_CompuMethod = 'Standard', the binary gas

diffusivities of any two components κ1 and κ2 depend on pressure and temperature as

[Vargaftik, 1975; Walker et al., 1981]

θ
κ1κ 2 κ1κ 2 P ! T +273.15 $
dG
( P,T ) = d G
( P0 ,T0 ) 0 # & (6.5)
P " 273.15 %

98
For example, at standard conditions of P0 = 1 atm = 1.01325 bar, T0 = 0 ˚C, the diffusion

coefficient for vapor-air mixtures has a value of 2.13x10-5 m2/s; the parameter θ for the

temperature dependence is 1.80. TOUGH+ v1.5 can model a temperature dependence of

gas phase diffusion coefficients according to Equation (6.5) by specifying parameter θ =

gas_diffusivity_equation_exponent in the first record of data block

DIFFUSION (see Section 6.4). Presently there are no provisions for allowing the

dependence of the parameter θ on the gas phase composition.

The computation of the effective diffusion coefficient in equation (2.60)

necessitates knowledge of the tortuosity. In TOUGH+ v1.5, this is computed using one of

the methods described by the variable Option_tortuosity_CompuMethod in data

block MEMORY (see Section 5.1).

6.5. Block-by-Block Permeability Modification

TOUGH+ v1.5 provides a feature that applies permeability modification (PM) coefficients

for individual grid blocks according to

kn → kn" = kn ⋅ ζ n (6.6)

Here, kn is the absolute (intrinsic or reference) permeability of grid block (cell,

element) n, as specified in data block ROCKS or MEDIA for the domain to which that grid

block belongs, while ζn is the permeability modification coefficient. The strength of

capillary pressure will be automatically scaled according to Leverett [1941] as:

99
 Pcap,n
Pcap,n → Pcap,n
" = (6.7)
ζn

The subroutine Initialize_Perm_Modifiers is called after all MESH data

have been processed. The Initialize_Perm_Modifiers routine initializes

permeability modifiers and generates informative printout. To engage block-by-block

permeability modification, users need to specify a (dummy) domain named "SEED" in

block ROCKS or MEDIA. No grid blocks should be assigned to this domain; the presence

of domain "SEED" simply serves as a flag to put permeability modification into effect,

while data provided in domain "SEED" serve to select different options. Random

(spatially uncorrelated) PM data can be internally generated in TOUGH+. Alternatively,

externally defined permeability modifiers may be provided as part of the geometry data in

block ELEME. This feature can be used to apply spatially correlated fields: users can run

their own geostatistical package to generate whatever fields they desire, and then use a

preprocessing program to place the PM-coefficients into the ELEME data block.

Note that this approach of block-by-block permeability modification affects only

grid block permeabilities but not porosities. Full details on the various user options for

PM-coefficients are given in informative printout that is automatically generated with each

TOUGH+ run.

In addition to this approach, it is possible for the users to apply a geostatistical

package of their choice to develop element-specific (block-by-block) statistically

heterogeneous porosities φ, intrinsic permeabilities k, and phase saturations. These data

can then be read as part of the INCON data block, as will be discussed in Section 8.

100
 7.0. Geometrical Representation,
Domain Discretization, and
Grid Generation

The data blocks that specify the geometrical representation of the simulated system are

discussed in this section, including the specification of elements (block ELEM),

connections between elements (block CONNE), and the generation of grids (block MESHM)

for radial symmetric grids, rectilinear grids, and grids containing fractured media. First,

we discuss the convention used in TOUGH+ v1.5 for entering and processing geometrical

data.

7.1. TOUGH+ Convention for Geometrical Data

Handling of flow geometry data in TOUGH+ is upward compatible with TOUGH2

[Pruess et al., 1999] and earlier TOUGH+ [Moridis et al., 2008] input formats and data

101
handling. As in other integral finite difference codes [Edwards, 1972; Narasimhan and

Witherspoon, 1976], flow geometry is defined by means of a list of volume elements (grid

blocks), and a list of flow connections between them (Section 3.3). This formulation can

handle regular and irregular flow geometries in one, two, and three dimensions. Single-

and multiple-porosity systems (porous and fractured media) can be specified, and higher

order methods, such as seven- and nine-point differencing, can be implemented by means

of appropriate specification of geometric data [Pruess and Bodvarsson, 1983].

In TOUGH+ v1.5, as in TOUGH2 [Pruess et al., 1999; 2012], volume elements

are identified by names that consist of a string of either five or eight characters, '12345'

or '12345678'. These are arbitrary, except that the last two characters (#4 and #5 in 5-

character names; #7 and #8 in 8-character names) must be numbers; examples of valid 5-

character and 8-character element names are 'ELE10' and 'AB00CC23', respectively.

Flow connections are specified as ordered pairs of elements, such as 'ELE10ELE11' (5-

character names) or 'AB00CC23AB00CC24' (8-character names). A variety of options

and facilities are available for entering and processing geometric data. As in TOUGH2

[Pruess et al., 1999; 2012], element volumes and domain identification can be provided

by means of a data block ELEME in the standard input file, while a data block CONNE can

be used to supply connection data, including interface area, nodal distances from the

interface, and orientation of the nodal line relative to the vertical (see Sections 7.2 and

7.3). These data are internally written to a disk file MESH, which in turn initializes the

geometry data arrays used during the flow simulation. It is also possible to omit the

ELEME and CONNE data blocks from the standard input file, and provide geometry data

directly in the MESH file.

102
 TOUGH+ v1.5 provides a capability for describing the flow system geometry and

discretizing the domain by means of the MeshMaker.f95 application written in

FORTRAN 95/2003 (see Section 7.4). Unlike the approach in TOUGH2 [Pruess et al.,

1999; 2012], MeshMaker.f95 has been separate from (i.e., it is not integrated within)

the core code since the first version of the TOUGH+ code [Moridis et al., 2008]. This

application can perform a number of mesh generation and processing operations. The

MeshMaker.f95 code is written according to the tenets of Object-Oriented

Programming, and has a modular structure. It can generate two-dimensional radially

symmetric (r,z) meshes, and one-, two-, and three-dimensional rectilinear (Cartesian) grids

in (x,y,z). Multiple-porosity processing for simulation of flow in naturally fractured

reservoirs can be invoked by means of a keyword MINC, which stands for Multiple

INteracting Continua (see Section 2.14). The MINC process operates on the data of the

primary (porous medium) mesh as provided on disk file MESH, and generates a secondary

mesh containing fracture and matrix elements with identical data formats on file MINC.

The file MESH used in this process can be either directly supplied by the user, or it can

have been generated from an earlier application of the MeshMaker.f95 application.

7.2. Data Block ELEME

This block introduces element (grid block) information.

Record ELEME.1 (for 5-character elements only)

Format (A3, I2, 2I5, A5, 6E10.4, 1X, A3)

ElName5C, NSEQ, NADD, MA12, elem_vol,

elem_aht, elem_pm, X, Y, Z, elem_activity

103
 ElName5C
The five-character name of an element. The first three characters can be
letters, numbers or blanks, and the last two characters (NE =
ElName5C(4:5)) must be numbers.
NSEQ
The number of additional elements having the same volume and belonging
to the same reservoir domain.

NADD
The increment between the code numbers of two successive elements.
(Note: if the longest dimension in the grid is <1000, the maximum
permissible code number NE + NSEQ *NADD is ≤ 999.)

MA12
A five-character material identifier corresponding to one of the reservoir
domains as specified in block ROCKS. If the first three characters are
blanks and the last two characters are numbers, then they indicate the
sequence number of the domain as entered in ROCKS. If left blank the
element is by default assigned to the first domain in block ROCKS.

elem_vol
The element volume (m3).

elem_aht
The interface area (m2) for heat exchange with semi-infinite confining
beds. Internal MESH generation via the MeshMaker.f95 facility will
automatically generate elem_aht.

elem_pm
The permeability modifier (optional, active only when a domain "SEED"
has been specified in the ROCKS block; see Section 6.5). Will be used as
multiplicative factor for the permeability parameters from block ROCKS.
Simultaneously, the capillary pressure strength will be scaled as elem_pm
-1/2
. elem_pm = 0 will result in an impermeable block.

Random permeability modifiers can be generated internally, see detailed

comments in the TOUGH+ output file. The elem_pm may be used to
specify spatially correlated heterogeneous fields, but users need their own
preprocessing programs for this, as TOUGH+ provides no internal
capabilities for generating such fields.

X, Y, Z
Cartesian coordinates of the grid block centers. These are included in the
ELEME data for listing in output files of the simulation results that conform
to the specifications of the TecPlot package. Additionally, the

104
 coordinate data may be needed for (a) the computation of interblock
properties in cylindrical grids or (b) the initialization of a gravity-capillary
equilibrium (such as the one described in Sections 6.9 and 9.10 of Pruess et
al. [1999]), and are needed (c) for mechanical dispersion computations in
solute transport investigations [Moridis et al., 1999].

elem_activity
A (optional) character variable that describes the activity status of an
element. An element is rendered inactive (i.e., a Dirichlet boundary) by
invoking the options elem_activity ='I__' (indicating that the
element conditions are time-invariant) or elem_activity = 'Vxx'
(denoting several time-variable conditions, each numbered by the xx
identifier). The grid block is assumed active by default for any other value
of elem_activity, including a blank (see Section 8.4).

Record ELEME.1 (for 8-character elements only)

Format (A8, 7X, A5, 6E10.4, 1X, A1)

ElName8C, MA12, elem_vol, elem_aht,

elem_pm, X, Y, Z, elem_activity

ElName8C
The eight-character name of an element. The first five characters can be
letters, numbers or blanks, and the last two characters (NE =
ElName8C(7:8)) must be numbers.

All other variables and parameters are as previously defined in case of the five-
character element names.

Repeat record ELEME.1 for the number of elements desired.

Record ELEME.2

A blank record closes the ELEME data block.

NOTE: The number of elements described in the ELEME data block cannot exceed
the number Max_NumElem specified in the MEMORY data block (See Section
5.1). If this happens, an error message is printed and the simulation is aborted.

105
7.3. Data Block CONNE

This block introduces information for the connections (interfaces) between elements.

Record CONNE.1 (for connections involving 5-character elements only)

Format (A5, A5, 4I5, 5E10.4)
ConxName1, ConxName2, NSEQ, NAD1, NAD2, ConxKi,
ConxD1, ConxD2, ConxArea, ConxBeta, emissivity
These parameters are defined as follows:
ConxName1
The name of the first five-character element in the connection.
ConxName2
The code name of the five-character second element in the connection.
NSEQ
The number of additional connections in the sequence.
NAD1
The increment of the number of the first element (as defined by the last two
characters of its name, i.e., ConxName1(4:5)) between two successive
connections.
NAD2
The increment of number of the second element (as defined by the last two
characters of its name, i.e., ConxName2(4:5)) between two successive
connections.
ConxKi
The permeability index of the connection. Setting it equal to 1, 2, or 3
specifies absolute permeability to be MediaPerm(ConxKi) for the
materials in elements (ConxName1) and (ConxName2), where
MediaPerm is read in block ROCKS. This allows assignment of different
permeabilities, e.g., in the horizontal and vertical direction.
ConxD1, ConxD2
The distance [m] between the common interface of a connection from the
centers of the first and second element in the connection, respectively.
ConxArea
The element interface area [m2].
ConxBeta
The cosine of the angle between the gravitational acceleration vector and
the line between the two elements. ConxBeta*gravity > 0 (<0)
corresponds to first element being above (below) the second element.

106
emissivity
The radiant emittance factor for radiative heat transfer, which for a
perfectly “black” body is equal to 1. The rate of radiative heat transfer
between the two grid blocks is:
QH ,rad = emissivity*σ 0 * ConxArea * (T24 − T14 )

where σ0 = 5.6687e-8 J/m2 K4 s is the Stefan-Boltzmann constant, and T1

and T2 are the absolute temperatures of the two grid blocks. The term
emissivity may be entered as a negative number, in which case the
absolute value will be used, and heat conduction at the connection will be
suppressed. Setting emissivity = 0.0e0 will result in no radiative
heat transfer.

Record CONNE.1 (for connections involving 5-character elements only)

Format (A8, A8, 9X, I5, 5E10.4)
Conx8Name1, Conx8Name2, ConxKi,
ConxD1, ConxD2, ConxArea, ConxBeta,
emissivity

Conx8Name1
The name of the first eight-character element in the connection.

Conx8Name2
The code name of the second eight-character element in the connection.

All other variables and parameters are as previously defined in case of connections
between five-character elements.

Repeat record CONNE.1 for the number of connections desired.

Record CONNE.2

A blank record closes the CONNE data block. Alternatively, connection

information may terminate on a record with '+++ ' typed in the first
five columns, followed by element cross-referencing information. The
second type of termination is generated in the MESH file upon completion
of a TOUGH+ run.

NOTE: The number of connections described in the CONNE data block cannot
exceed the number Max_NumConx specified in the MEMORY data block (See
Section 5.1). If this happens, an error message is printed and the simulation is
aborted

107
7.4. The MeshMaker.f95 Facility

In this section we discuss the use of the MeshMaker.f95 facility, and the required

parameter inputs for mesh generation and processing. As indicated earlier, unlike in

TOUGH2 [Pruess et al., 1999; 2012], MeshMaker.f95 is not integrated into the

TOUGH+ code but is an independent program that is written in FORTRAN 95/2003, has

a modular structure, and an architecture based on the principles of Object-Oriented

Programming. The MeshMaker.f95 input has a modular structure, which is organized

by keywords in a manner analogous to that of TOUGH+. This section provides detailed

instructions for preparing the input files, illustrative examples of which are shown in

Figures 7.1 to 7.4.

7.4.1. Inputs Related to Problem Definition and Dimensioning

These inputs occupy two records (both mandatory), and provide (a) a short description of

the problem through an informative title, and (b) data to allow proper dimensioning of the

work arrays and formatting of the MESH file to be generated. These two initial records are

discussed in detail below:

Record MESHMAKER.1

The first record of the input file in any MeshMaker.f95 application is the
character variable TITLE, which includes a header of up to 80 characters and is
read using a free format. This record is necessary for any MeshMaker simulation
to begin.

Record MESHMAKER.2

The following variables are read in MESHMAKER.2 using a free format:

108
 MaxNum_Elem, Longest, ElemNameLength,
FormatType, LengthUnits, media_by_number

These parameters are defined as follows:

Max_NumElem
Integer denoting the maximum number of elements in the grid under
construction.

Longest
Integer indicating the maximum expected number of subdivisions along
any of the coordinates in the grid under construction.
ElemNameLength
Integer variable defining the number of characters in the element names. It
may be either 5 or 8. If unequal to 8, ElemNameLength is internally
resest to the default (= 5).
FormatType
Character variable indicating the format of the data in the MESH file to be
created. It may have one of two values:
= 'Old': This option creates a MESH file that conforms to the data format
described in Sections 7.2 and 7.3, and is consistent with the TOUGH2
formats [Pruess et al., 1999; 2012].
= 'New': This option creates a MESH file in which the element and
connection data are listed using the NAMELIST format facility available in
FORTRAN 95/2003. This is intended for future versions of TOUGH+
NOTE: TOUGH+ v1.5 does not accept NAMELIST-based formats in the
MESH file. Thus, the option FormatType = 'Old' must be used in all
MeshMaker.f95 applications.

LengthUnits
Character variable specifying the units of length of the grid to be created.
The possible options are: 'm', 'mm', 'km', 'ft' and 'in', indicating meters,
millimeteres, kilometers, feet and inches, respectively. Note that all
dimensions are converted internally into SI units (meters), which is the
standard unit of length of the resulting mesh.

media_by_number
Logical variable specifying how the various media (rocks) are to be named.
The possible options are: .T. or .TRUE., in which case the media are named
by the character variable H_RegionName (see Section 7.4.2), and .F. of
.FALSE., indicating media identification by the number in the sequence of
their listing. The media_by_number = .T. option is to be used only in
the case of fractured media that are to be described by the Multiple

109
 Intercative Continua (MINC) concept, and is needed for the creation of an
MESH interim files. This will be further processed to create the MINC file
to be used in the simulation. When the media_by_number = .TRUE.
option is invoked, unfractured media are identified by a negative number,
and fractured media (identified by an ‘F’ or ‘f’ as the first letter of the value
of the H_RegionName character variable) are identified by a positive
number.

7.4.2. Inputs Related to Domain Heterogeneity

These optional inputs provide information that allows the description of heterogeneity

within the domain. The assignment of heterogeneity is based on the definition of regions

using geometrical information that describe the location and extent of these subdomains.

Thus, each of the individually defined regions is assigned the properties of a particular

medium that will have to be included in the ROCKS data block in the input file of the

subsequent TOUGH+ v1.5 simulation.

The records in this data block are discussed in detail below:

Record MESHMAKER.3 (Optional)

This record includes only the keyword (character variable) 'Regions', which is
read using a free format. If the domain is homogeneous, then there is no need to
provide any of the inputs discussed in Section 7.4.2.
Record MESHMAKER.3.1 (Optional – when the MESHMAKER.3 record is
included)

This record is read using a free format, and includes the single integer variable
Num_HetRegions (> 0) that describes the number of heterogeneous regions
(subdomains) that are to be described by the ensuing data records. The minimum
value that Num_HetRegions can accept is 1, corresponding to a homogeneous
system. An error will occur and the execution will stop if Num_HetRegions <
1.

Because of dynamic dimensioning, there is no limit in the number of

heterogeneous regions defined by Num_HetRegions. However, practical
considerations may limit the size of Num_HetRegions. Thus, although it is
possible to define element-by-element heterogeneity using this approach, this
would be a very tedious process. Generally speaking, the most useful application

110
of this facility is in the description of extensive geologic units with distinctly
different properties.

Record MESHMAKER.3.2 (Optional – when the MESHMAKER.3 record is

included)

This record is read using a free format, and includes the single character variable
dominant_medium that provides the name (5 character long) of the dominant
(reference) porous medium in the heterogeneous domain. Note that the selection
of a medium as dominant_medium is be arbitrary, as there are no restrictions
on the extent of its spatial distribution for it to be designated as such.

If Num_HetRegions = 1, no more data need to be read. This case is equivalent

to a homogeneous system, and the entire heterogeneity-related data block (records
MESHMAKER.3 to MESHMAKER.3.2 may be omitted.

Record MESHMAKER.3.3.0 (Optional – when the MESHMAKER.3 record is

included and Num_HetRegions > 1)

This record is read using a free format, and includes the single character variable
H_RegionName that provides the name (5 character long) of the porous medium
in the region (subdomain) that is about to be defined.

Record MESHMAKER.3.3.1 (Optional – when the MESHMAKER.3 record is

included and Num_HetRegions > 1)

Here the following character variables are read using a free format:
H_RegionCoordinates, H_RegionUnits
These parameters are defined as follows:

H_RegionCoordinates
A character variable indicating the coordinate system used in the geometric
definition of the region that is about to be described in the heterogeneous
domain. It my have one of two values:

= 'Cartesian' : This option indicates that the region is defined

geometrically in terms of Cartesian coordinates.

= 'Cylindrical' : This option indicates that the region is defined

geometrically in terms of cylindrical coordinates.

H_RegionUnits
A character variable describing the units of length used in the description
of the geometry of the region. The following values are acceptable options:

111
 'mm', 'km', 'km', 'in', or 'ft',
indicating millimeters, meters, kilometers, inches and feet, respectively.
Note that MeshMaker.f95 converts all length units into meters (the
length unit used in the TOUGH+ simulations) prior to producing the MESH
file.

Record MESHMAKER.3.3.2 (Optional – when the MESHMAKER.3 record is

included and Num_HetRegions > 1)

If H_RegionCoordinates = 'Cartesian', the following real variables are

read in MESHMAKER.3.3.2 using a free format:
Xmin, Xmax, Ymin, Ymax, Zmin, Zmax
These parameters are defined as follows:

Xmin, Xmax
Real variables indicating the range of the region along the x-axis of the
Cartesian coordinate system.

Ymin, Ymax
Real variables indicating the range of the region along the y-axis of the
Cartesian coordinate system.
Zmin, Zmax
Real variables indicating the range of the region along the z-axis of the
Cartesian coordinate system.

If H_RegionCoordinates = 'Cylindrical', the following real variables

are read in MESHMAKER.3.3.2 using a free format:

Rmin, Rmax, Zmin, Zmax

These parameters are defined as follows:

Rmin, Rmax
Real variables indicating the range of the region along the r-axis of the
cylindrical coordinate system.

Zmin, Zmax
Real variables indicating the range of the region along the z-axis of the
cylindrical coordinate system.

Repeat records MESHMAKER.3.3.0, MESHMAKER.3.3.1,

MESHMAKER.3.3.3 for a total of Num_HetRegions – 1 regions. Note that
the region dominant_medium does not need geometric definition.

112
7.4.3. Inputs Related to Description of Boundaries

These optional inputs provide information that describes the outer boundaries of the

domain under discretization. The assignment of grid subdomains as boundaries is based

on the geometry-based definition of the outer spatial limits of the domain under

discretization. Thus, the boundaries are treated as special types of regions (see Section

7.4.2) with the properties of a particular medium that will have to be included in the

ROCKS data block in the input file of the subsequent TOUGH+ simulation. It is possible

for a region and one or more boundaries to be described by the same medium in the

ROCKS data block.

The records in this data block are discussed in detail below:

Record MESHMAKER.4 (Optional)

This record includes only the keyword (character variable) 'Boundaries', which
is read using a free format. If the domain is confined by no-flow (Newman-type)
boundaries, then there is no need to provide any of the inputs discussed in Section
7.4.3.

Record MESHMAKER.4.1 (Optional – when the MESHMAKER.4 record is

included)

This record is read using a free format, and includes the single integer variable
Num_Boundaries (> 0) that describes the number of boundaries that are to be
described.

Record MESHMAKER.4.2.0 (Optional – when the MESHMAKER.4 record is

included and Num_Boundaries > 1)
The following character variables are read in MESHMAKER.4.2 using a free
format:
BoundID, BoundRegionName
These parameters are defined as follows:

113
 BoundID
A character variable indicating the type of boundary described at the
prescribed location. It my have one of two values:
= 'I': This option indicates an inactive boundary, the conditions and
properties of which are time-invariant.
= 'V': This option indicates a time-variable boundary.
The result of this designation is reflected in the ELEME block, in which
elem_activity (see Section 7.2) is set to BoundID in all the cells
corresponding to the boundary defined here. Any other value of the
BoundID variable causes the program to print an error message and stop
execution.
BoundRegionName
A character variable H_RegionNname that provides the name (5
character long) of the porous medium in the boundary that is about to be
defined.

Record MESHMAKER.4.2.1 (Optional – when the MESHMAKER.4 record is

included and Num_Boundaries > 1)

Here the following character variables are read using a free format:
BoundRegionCoordinates, BoundRegionUnits
These parameters are defined as follows:

BoundRegionCoordinates
A character variable indicating the coordinate system used in the geometric
definition of the boundary that is about to be described in the
heterogeneous domain. It my have one of two values:
= 'Cartesian': This option indicates that the region is defined
geometrically in terms of Cartesian coordinates.

= 'Cylindrical': This option indicates that the region is defined

geometrically in terms of cylindrical coordinates.

BoundRegionUnits
A character variable describing the units of length used in the description
of the geometry of the boundary. The following values are acceptable
options:
'mm', 'km', 'km', 'in', or 'ft',
indicating millimeters, meters, kilometers, inches and feet, respectively.
Note that MeshMaker.f95 converts all length units into meters (the

114
 length unit used in the TOUGH+ simulations) prior to producing the MESH
file.

Record MESHMAKER.4.2.2 (Optional – when the MESHMAKER.4 record is

included and Num_Boundaries > 1)

If BoundRegionCoordinates = 'Cartesian', the following real variables

are read in MESHMAKER.3.3.2 using a free format:

Xmin, Xmax, Ymin, Ymax, Zmin, Zmax

These parameters are defined as follows:

Xmin, Xmax
Real variables indicating the range of the boundary region along the x-axis
of the Cartesian coordinate system.

Ymin, Ymax
Real variables indicating the range of the boundary region along the y-axis
of the Cartesian coordinate system.

Zmin, Zmax
Real variables indicating the range of the boundary region along the z-axis
of the Cartesian coordinate system.

If BoundRegionCoordinates = 'Cylindrical', the following real

variables are read in MESHMAKER.3.3.2 using a free format:
Rmin, Rmax, Zmin, Zmax
These parameters are defined as follows:

Rmin, Rmax
Real variables indicating the range of the region boundary along the r-axis
of the cylindrical coordinate system.

Zmin, Zmax
Real variables indicating the range of the boundary region along the z-axis
of the cylindrical coordinate system.

Repeat records MESHMAKER.4.2.0, MESHMAKER.4.2.1,

MESHMAKER.4.2.3 for a total of Num_Boundaries boudaries.

115
7.4.4. Inputs for Grid Construction

There are three grid construction options available in MeshMaker.f95. These options

are activated by appropriate keywords. Thus, the keywords 'RZ2D' or 'RZ2DL' invoke

generation of a one or two-dimensional radially symmetric (r,z) mesh; 'XYZ' initiates

generation of a one, two, or three dimensional Cartesian (x,y,z) mesh; and 'MINC' calls a

modified version of the GMINC program [Pruess, 1983] to sub-partition a primary porous

medium mesh into a secondary mesh for fractured media, using the method of Multiple

INteracting Continua [Pruess and Narasimhan, 1982; 1985].

The meshes generated under keyword 'RZ2D' or 'XYZ' are internally written to file

MESH. The MINC-processing operates on the data in a file MESH that is either pre-

existing or is created directly from the inputs of MeshMaker.f95 when the 'RZ2D',

'RZ2DL' or 'XYZ' data blocks are followed by the 'MINC' datablock. We shall now

separately describe the preparation of input data for the three grid construction options.

7.4.4.1. Generation of radially symmetric grids (keyword 'RZ2D' or 'RZ2DL'). The

keywords 'RZ2D' or 'RZ2DL' invoke generation of a radially symmetric mesh. Values for

the radii to which the grid blocks extend can be provided by the user or can be generated

internally (see below). Nodal points will be placed halfway between neighboring radial

interfaces. When 'RZ2D' is specified, the mesh will be generated by columns; i.e., in the

ELEME block we will first have the grid blocks at smallest radius for all layers, then the

next largest radius for all layers, and so on.

With keyword 'RZ2DL' the mesh will be generated by layers; i.e., in the ELEME

block we will first have all grid blocks for the first (top) layer from smallest to largest

radius, then all grid blocks for the second layer, and so on. Apart from the different

116
ordering of elements, the two meshes for 'RZ2D' and 'RZ2DL' are identical. The reason for

providing the two alternatives is as a convenience to users in implementing boundary

conditions by way of inactive elements (see Section 6.4 in Pruess et al. [1991]).

Assignment of inactive elements would be made by using a text editor on the RZ2D-

generated MESH file, and moving groups of elements towards the end of the ELEME block,

past a dummy element with zero volume. 'RZ2D' makes it easy to declare a vertical

column inactive, facilitating assignment of boundary conditions in the vertical, such as a

gravitationally equilibrated pressure gradient. 'RZ2DL' on the other hand facilitates

implementation of areal (top and bottom layer) boundary conditions.

The inputs for cylindrical systems are as follows:

Data Block GRID

The first record in this data block includes only the keyword (character variable)
'RZ2D' or 'RZ2DL'. This keyword is read using a Format (A5).

Record RADII.0

RADII: A keyword that introduces data for defining a set of interfaces (grid
block boundaries) in the radial direction. It is read using a Format (A5).

Record RADII.1
Format (I5)
NRAD

NRAD: Number of radii that will be read. At least one radius must be provided,
indicating the inner boundary of the mesh.

Record RADII.2, RADII.3, etc.

Format (8E10.4)
RC(i), i = 1,NRAD

RC(i): The radii defining the element boundaries in ascending order [m].

117
 Record EQUID.0

EQUIDistant
Keyword indicating that the ensuing data describe a set of equal radial
increments. This keyword is read using Format (A5).

Record EQUID.l
Format (I5, 5X, E10.4)
NEQU, DR
NEQU: The number of desired radial increments.
DR: The size of the radial increment [m].
NOTE: At least one radius must have been defined via block RADII before
EQUID can be invoked.

Record LOGAR.0
LOGARithmic This keyword introduces data on radial increments
that increase from one to the next by the same factor
(i.e., Δrn+1 = f • Δrn).

Record LOGAR.l
Format(I5, 5X, 2E10.4)
NLOG, RLOG, DR
NLOG: The number of desired radial increments.
RLOG: The desired radius rmax of the last (largest) of these radii.

DR: The reference radial increment Δr0: the first Δr generated will be equal to
f • Δr0, with f internally determined such that the last increment will bring
total radius to RLOG = rmax. The factor f < 1 for decreasing radial
increments is permissible. If Δr0 is set equal to zero, or left blank, the last
increment Δr generated before the keyword LOGAR is invoked will be
used as default.
Additional blocks RADII, EQUID, and LOGAR can be specified in arbitrary order.
NOTE: At least one radius must have been defined before the LOGAR option can
be invoked. If Δr0 = 0, at least two radii must have been defined.

118
 Record LAYER.0
LAYER: This keyword introduces information on horizontal layers, and signals
closure of RZ2D input data. It is read using a Format (A5).
Record LAYER.l
Format (I5)
NLAY

NLAY: The number of horizontal layers in the cylindrical grid.

Record LAYER.2
Format (8E10.4)
H(i), i = 1, NLAY
H(i): The thicknesses of the horizontal layers in the cylindrical grid, from top
layer downward. By default, zero or blank entries for layer thickness will
result in assignment of the last preceding non-zero entry. Assignment of
a zero layer thickness, as needed for inactive layers, can be accomplished
by specifying a negative value.

NOTE: The LAYER data close the RZ2D or RZ2DL data block.

Record GRID-END
Two blank records close the input data file if further MINC-processing of the grid
is not needed. If such processing is specified, then a single blank record terminates
the RZ2D or RZ2DL data block, and is then followed by the MINC data block (see
Section 7.2.5).

7.4.4.2. Generation of rectilinear grids (keyword 'XYZ')

Data Block GRID

The first record in this data block includes only the keyword (character variable)
'XYZ' that is invokes generation of a Cartesian (rectilinear) mesh. This keyword is
read using a Format (A5).

Record XYZ.l
DEG, X_ref, Y_ref, Z_ref (Free format)
DEG: The angle (in degrees) between the y-axis and the horizontal. If
gravitational acceleration (parameter gravity in record PARAM.2, see
Section 10) is positive, -90° < DEG < 90° corresponds to grid layers
going from top down. Grids can be specified from the bottom layer up

119
 by setting gravity or ConxBeta (Section 7.3) to negative values.
The default (DEG = 0.0e0) corresponds to horizontal y- and vertical z-
axis. The x-axis is always horizontal.
X_ref, Y_ref, Z_ref:
The reference x-, y-, and z-coordinates at the origin of the axes.

Record XYZ.2
Format (A2, 3X, I5, E10.4)
NTYPE, NO, DEL
NTYPE: A character variable that can assume one of the values 'NX', 'NY' or 'NZ',
specifying grid increments in the x-, y-, or z-direction, respectively.
NO: The number of grid increments.
DEL: The constant grid increment for NO grid blocks, if set to a non-zero value.

Record XYZ.3 (Optional, DEL = 0.0e0 or blank only)

Format (8E10.4)
DEL(i), i = 1, NO
DEL(i): A set of grid increments in the direction specified by NTYPE in record
XYZ.2. Additional records with formats as XYZ.2 and XYZ.3 can be
provided, with x-, y-, and z-data in arbitrary order.

Record XYZ.4 A blank record closes the XYZ data block.

NOTE: The end of the data block GRID is also marked by a blank record. Thus,
when GRID/XYZ is used, there will be two blank records at the end of the input
data file. Alternatively, a single blank record and a new record with the keyword
'MINC' can be used to invoke MINC-processing for fractured media (see below).

120
 Meshmaker test: Cartesian grid
1000 20 5 'Old' 'm' .FALSE.
XYZ
00.
NX 10 1.0e00
NY 5 2.0
NZ 20 1.0

Figure 7.1. An example of a MeshMaker.f95 input file for the creation of a Cartesian 3D grid.
Note that no heterogeneous regions or boundaries are defined in this grid.

Meshmaker test: Cylindrical grid

6000 6000 5 'Old' 'm' .FALSE.
RZ2DL
RADII
2
1.0e-5 1.0e-3
EQUID
4000 2.5e-2
LOGAR
500 2.5e+3
EQUID
1 1.0e-3
LAYER
1
1.0e00

Figure 7.2. An example of a MeshMaker.f95 input file for the creation of a single-layer (1D)
cylindrical grid. Note that no heterogeneous regions or boundaries are defined in this grid

121
Input file for a large 3D cartesian grid ! Title
500000 100 5 'Old' 'm' .FALSE.
Regions ! Keyword denoting heterogeneous subdomains
6 ! ==> # of heterogeneous media regions
'HydrL' ! Region #1: Name of dominant medium
'Aquif' ! Region #2: medium name
'cartesian' 'm' ! coordinates, units
0.0e0 1.5e3 0.0e0 1.5e3 -6.3e1 -4.825e1 ! Xmin, Xmax, Ymin, Ymax, Zmin, Zmax
'OverB' ! Region #3: medium name
'cartesian' 'm' ! coordinates, units
0.0e0 1.5e3 0.0e0 1.5e3 -3.0e1 0.0e1 ! Rmin, Rmax, Zmin, Zmax
'UndrB' ! Region #4: medium name
'cartesian' 'm' ! coordinates, units
0.0e0 1.5e3 0.0e0 1.5e3 -9.4e1 -6.3e1 ! Rmin, Rmax, Zmin, Zmax
'WellA' ! Region #5: medium name
'cartesian' 'm' ! coordinates, units
0.0e0 5.0e-2 0.0e0 5.0e-2 -5.4e1 -3.0e1 ! Rmin, Rmax, Zmin, Zmax
'WellA' ! Region #6: medium name
'cartesian' 'm' ! coordinates, units
4.9995e2 5.0e2 4.9995e2 5.0e2 -5.4e1 -3.0e1 ! Rmin, Rmax, Zmin, Zmax
Boundaries
2 ! ==> # of boundaries
'I' 'TopBB' ! Boundary #1: type and medium name
'cartesian' 'm' ! coordinates, units
0.0e0 1.5e3 0.0e0 1.5e3 -1.0e-2 0.0e0 ! Rmin, Rmax, Zmin, Zmax
'I' 'BotBB' ! Boundary #2: type and medium name
'cartesian' 'm' ! coordinates, units
0.0e0 1.5e3 0.0e0 1.5e3 -9.4e1 -9.3e1 ! Rmin, Rmax, Zmin, Zmax
XYZ
00.
NX 76
5.0e-2 2.5000e-1 2.8599e-1 3.2716e-1 3.7426e-1 4.2813e-1 4.8976e-1 5.6027e-1
6.4092e-1 7.3319e-1 8.3873e-1 9.5947e-1 1.0976e+0 1.2556e+0 1.4363e+0 1.6431e+0
1.8797e+0 2.1502e+0 2.4598e+0 2.8139e+0 3.2190e+0 3.6823e+0 4.2124e+0 4.8188e+0
5.5125e+0 6.3061e+0 7.2139e+0 8.2524e+0 9.4404e+0 1.0799e+1 1.2354e+1 1.4132e+1
1.6167e+1 1.8494e+1 2.1157e+1 2.4202e+1 2.7686e+1 3.1672e+1 3.1672e+1 2.7686e+1
2.4202e+1 2.1157e+1 1.8494e+1 1.6167e+1 1.4132e+1 1.2354e+1 1.0799e+1 9.4404e+0
8.2524e+0 7.2139e+0 6.3061e+0 5.5125e+0 4.8188e+0 4.2124e+0 3.6823e+0 3.2190e+0
2.8139e+0 2.4598e+0 2.1502e+0 1.8797e+0 1.6431e+0 1.4363e+0 1.2556e+0 1.0976e+0
9.5947e-1 8.3873e-1 7.3319e-1 6.4092e-1 5.6027e-1 4.8976e-1 4.2813e-1 3.7426e-1
3.2716e-1 2.8599e-1 2.5000e-1 5.0e-2
NY 76
5.0e-2 2.5000e-1 2.8599e-1 3.2716e-1 3.7426e-1 4.2813e-1 4.8976e-1 5.6027e-1
6.4092e-1 7.3319e-1 8.3873e-1 9.5947e-1 1.0976e+0 1.2556e+0 1.4363e+0 1.6431e+0
1.8797e+0 2.1502e+0 2.4598e+0 2.8139e+0 3.2190e+0 3.6823e+0 4.2124e+0 4.8188e+0
5.5125e+0 6.3061e+0 7.2139e+0 8.2524e+0 9.4404e+0 1.0799e+1 1.2354e+1 1.4132e+1
1.6167e+1 1.8494e+1 2.1157e+1 2.4202e+1 2.7686e+1 3.1672e+1 3.1672e+1 2.7686e+1
2.4202e+1 2.1157e+1 1.8494e+1 1.6167e+1 1.4132e+1 1.2354e+1 1.0799e+1 9.4404e+0
8.2524e+0 7.2139e+0 6.3061e+0 5.5125e+0 4.8188e+0 4.2124e+0 3.6823e+0 3.2190e+0
2.8139e+0 2.4598e+0 2.1502e+0 1.8797e+0 1.6431e+0 1.4363e+0 1.2556e+0 1.0976e+0
9.5947e-1 8.3873e-1 7.3319e-1 6.4092e-1 5.6027e-1 4.8976e-1 4.2813e-1 3.7426e-1
3.2716e-1 2.8599e-1 2.5000e-1 5.0e-2
NZ 77
1.0e-3 7.00e0 5.00e0 4.0e+0 3.2e+0 2.5e+0 2.0e00 1.6e00
1.25e00 1.0e00 8.0e-1 6.5e-1 5.0e-1 5.0e-1 4.0e-1 4.0e-1
4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1
4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1
4.0e-1 4.0e-1 4.5e-1 4.5e-1 4.5e-1 4.5e-1 4.5e-1 4.0e-1
4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1
4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1 4.0e-1
4.0e-1 4.0e-1 4.0e-1 5.0e-1 5.0e-1 6.5e-1 8.0e-1 1.0e00
1.0e00 1.25e00 1.6e00 2.0e00 2.5e00 3.2e00 3.0e00 4.00e0
4.0e00 5.0e00 6.0e00 8.0e00 1.0e-3

====> ====> ====> ====>

Figure 7.3. An example of a MeshMaker.f95 input file for the creation of a large Cartesian 3D
grid with heterogeneous regions and defined boundaries.

122
Input file for a large cylindrical grid ……! Title
15000 30000 5 'Old' 'm' .FALSE.
Regions ! Keyword denoting heterogeneous subdomains
6 ! => Num_HetRegions ( = number of heterogeneous regions)
'HydrL' ! Region #1: dominant_medium
'Aquif' ! Region #2: H_RegionName
'cylindrical' 'm' ! H_RegionCoordinates, H_RegionUnits
0.0e0 1.5e3 -6.3e1 -4.825e1 ! Rmin, Rmax, Zmin, Zmax
'OverB' ! Region #3: H_RegionName
'cylindrical' 'm' ! H_RegionCoordinates, H_RegionUnits
0.0e0 1.5e3 -3.0e1 0.0e1 ! Rmin, Rmax, Zmin, Zmax
'UndrB' ! Region #4: H_RegionName
'cylindrical' 'm' ! H_RegionCoordinates, H_RegionUnits
0.0e0 1.5e3 -9.4e1 -6.3e1 ! Rmin, Rmax, Zmin, Zmax
'Casng' ! Region #5: H_RegionName
'cylindrical' 'm ' ! H_RegionCoordinates, H_RegionUnits
0.0e0 1.0e-1 -4.6e1 -3.0e1 ! Rmin, Rmax, Zmin, Zmax
'Perfo' ! Region #5: H_RegionName
'cylindrical' 'm ' ! H_RegionCoordinates, H_RegionUnits
0.0e0 1.0e-1 -5.225e1 -4.6e1 ! Rmin, Rmax, Zmin, Zmax
Boundaries ! Keyword denoting boundaries
2 ! ==> Num_Boundaries
'I' 'TopBB' ! Boundary #1: BoundID, BoundRegionName
'cylindrical' 'm' ! BoundRegionCoordinates, BoundRegionUnits
0.0e0 1.5e3 -1.0e-2 0.0e0 ! Rmin, Rmax, Zmin, Zmax
'I' 'BotBB' ! Boundary #2: BoundID, BoundRegionName
'cylindrical' 'm' ! BoundRegionCoordinates, BoundRegionUnits
0.0e0 1.5e3 -9.4e1 -9.3e1 ! Rmin, Rmax, Zmin, Zmax
RZ2DL
RADII
2
0.001 0.10795
EQUID
1 2.0e-2
LOGAR
199 1.0e+3
LAYER
68
1.0e-3 6.50e0 5.00e0 5.0e+0 3.0e+0 3.0e+0 2.0e00 2.0e00
1.0e00 1.0e00 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1
5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1
5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1
5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1
5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.5e-1 5.5e-1 5.5e-1 5.5e-1
5.5e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 5.0e-1 1.0e00
1.0e00 2.0e00 2.0e00 3.0e+0 3.0e+0 4.0e+0 4.0e+0 5.00e0
5.0e+0 6.00e0 6.0e+0 1.0e-3

====> ====> ====> ====>

Figure 7.4. An example of a MeshMaker.f95 input file for the creation of a large cylindrical 2D
grid with multiple layers, heterogeneous regions and defined boundaries.

123
7.4.4.3. MINC processing for fractured media (keyword 'MINC')

Data Block GRID

The first record in this data block includes only the keyword (character variable)
'MINC' that is invokes post-processing of a primary porous medium mesh from a
previously developed file MESH. This keyword is read using a Format (A5). The
input formats in data block MINC are identical to those of the GMINC program
[Pruess, 1983], with two enhancements: (a) there is an additional facility for
specifying global matrix-matrix connections (dual permeability option); (b) only
active elements (see Section 8.4) will be subjected to MINC-processing, the
remainder of the MESH remaining unaltered as porous medium grid blocks. See
Section 2.14 for further discussion.

NOTE: If the application of the MeshMaker.f95 aims to process a pre-existing

external MESH file, the input file begins with records MESHMAKER.1 and
MESHMAKER.2 (see Section 7.2.1) as the first two data blocks, followed by the
data inputs described in detail below.

Record MINC.1
Format (2A5, 5X, A5)
PART, TYPE, DUAL
PART: This is the first keyword following the 'MINC' keyword. It will be
followed on the same line by parameters TYPE and DUAL with
information on the nature of fracture distributions and matrix-matrix
connections.

PART: An identifier of the data block with partitioning parameters for secondary
mesh.

TYPE: A five-character variable for selecting one of the following six different
proximity functions provided in MINC [Pruess, 1983].
= 'ONE-D': A set of plane parallel infinite fractures with matrix block
thickness between neighboring fractures equal to PAR(l).
= 'TWO-D': Two sets of plane parallel infinite fractures, with arbitrary
angle between them. Matrix block thickness is PAR(l) for the first set,
and PAR(2) for the second set. If PAR(2) is not specified explicitly, it
will be set equal to PAR(l).
= 'THRED': Three sets of plane parallel infinite fractures at right
angles, with matrix block dimensions of PAR(l), PAR(2), and
PAR(3), respectively. If PAR(2) and/or PAR(3) are not explicitly
specified, they will be set equal to PAR(l) or PAR(2), respectively.

124
 = 'STANA': Average proximity function for rock loading of Stanford
large reservoir model [Lam et al., 1988].

= 'STANB': Proximity function for the five bottom layers of Stanford

large reservoir model.

= 'STANT': Proximity function for top layer of Stanford large reservoir

model.

DUAL A five-character word for selecting the treatment of global matrix flow.

= ' ' (Blank – default): The global flow occurs only through the
fracture continuum, while rock matrix and fractures interact locally by
means of interporosity flow (double-porosity model).

= 'MMVER': The global matrix-matrix flow is permitted only in the

vertical; otherwise like the double-porosity model; for internal
consistency this choice should only be made for flow systems with one or
two predominantly vertical fracture sets.

= 'MMALL': The global matrix-matrix flow in all directions; for internal

consistency only two continua, representing matrix and fractures, should
be specified (dual-permeability model).

NOTE: A user wishing to employ a different proximity function other than the
options provided through the TYPE variable in MINC needs to replace the function
subprogram PROX(x) in MeshMaker.f95 with a routine of the form:

FUNCTION PROX(x)
PROX = (arithmetic expression in x)
RETURN
END

It is necessary that PROX(x) be defined even when x exceeds the maximum

possible distance from the fractures, in which case PROX = 1. Additionally,
when the users supply their own proximity function subprogram, the parameter
TYPE must be set equal to 'ONE-D', 'TWO-D', or 'THRED', depending on the
dimensionality of the proximity function. This will assure proper definition of the
innermost nodal distances [Pruess, 1983].

Record PART.l
Format (2I3, A4, 7E10.4)
J, NVOL, WHERE, (PAR(i), i = 1, 7)

J: The total number of multiple interacting continua (J < 36).

125
 NVOL: The total number of explicitly provided volume fractions (NVOL < J).
If NVOL < J, the volume fractions with indices NVOL+1, ..., J will be
internally generated; all being equal and chosen such as to yield proper
normalization to 1.

WHERE: Character variable specifying whether the sequentially specified volume

fractions begin with the fractures (WHERE = 'OUT') or in the interior
of the matrix blocks (WHERE = 'IN').

PAR(i): Real array that stores the parameters describing the fracture spacing (see
discussion of previous record).

Record PART.2.1, PART.2.2, etc.

Format (8E10.4)
(VOL(i), i = 1, NVOL)

VOL(i): The volume fraction (having a value between 0 and 1) of a continuum

with index i (for WHERE = 'OUT') or index J+l-i (for WHERE =
'IN'). NVOL volume fractions will be read. For WHERE = 'OUT',
i=1 is the fracture continuum, i=2 is the matrix continuum closest to
the fractures, i=3 is the matrix continuum adjacent to i=2, etc. The sum
of all volume fractions must not exceed 1.

Record GRID-END
Two blank records close the input data file.

126
 8.0. Initial Conditions and
Boundary Conditions

In this section the data blocks that allow for domain-specific initial conditions (block

INDOM), element-specific initial conditions (INCON), extended capabilities for specifying

initial conditions (EXT-INCON), including features for assigning initial conditions by a

variety of methods. Following that the procedure is described for implementing initial

conditions and boundary conditions in TOUGH+ v1.5.

8.1. Data Block INCON

This block introduces element-specific initial conditions.

Record INCON.1 (for 5-character element names)

Format (A5, 2I5, E15.8, 2x, A3, 36x, 3(E15.8))
ElName5C, NSEQ, NADD, porosity, StateIndex, (perm(i), i=1,3)

127
 ElName5C
The 5-character name of the element that is being initialized.

NSEQ
The number of additional elements with the same initial conditions.
NADD
The increment between the code numbers of two successive elements with
identical initial conditions.
Porosity
The porosity of the element that is being initialized. If porosity is zero
or blank, the element porosity will be taken as specified in block ROCKS or
MEDIA. This feature is necessary for assignment of element-specific
properties for the description of highly or statistically heterogeneous
domains.
StateIndex
State identifier (see Section 3.1, Tables 3.1 and 3.2): the initial conditions
corresponding to this identifier are applied uniformly over the element.
perm(i), i=1,3
The intrinsic (absolute) permeabilities of the element that is being
initialized along the three directions described by the ConxKi variable
(see Section 7.3). If all perm(i), i=1,…,3 are zero or blank, the
element permeabilities will be taken as specified in block ROCKS or
MEDIA. This feature is necessary for assignment of element-specific
properties for the description of highly or statistically heterogeneous
domains.

Record INCON.1 (for 8-character element names)

Format (A8, 7X, E15.8, 2x, A3, 36x, 3(E15.8))

ElName8C, NSEQ, NADD, porosity, StateIndex, (perm(i), i=1,3)

ElName8C is the 8-character name of the element that is being initialized. All
other variables remain as in the case of five-character element names.

Record INCON.2 specifies primary variables.

Format (6E20.13)
X(i), i = 1,NumCom+1

The primary variables defining the state of the element specified in record
INCON.l. INCON specifications will supersede default conditions specified in

128
 PARAM.4 (see Section 10), and domain-specific conditions that may have been
specified in data block INDOM. See Section 3.1 (Tables 3.1 and 3.2) for a detailed
description of the potential sets of primary variables.

Record INCON.3

There are three ways to close the INCON data block.

If the simulation is not a continuation run, then a blank line or a record with ‘<<<’
typed in the first three columns closes the INCON data block.

For continuation runs from a previous TOUGH+ v1.5 simulation, the element-
related INCON data terminate with a record with ‘:::’ typed in the first three
columns. This is followed by a namelist that includes data describing the origin of
time, the simulated time at the conclusion of the preceding TOUGH+ v1.5
simulation and the cumulative numbers of the timesteps and of the Newtonian
iterations at the end of the previous run (see Figure 8.1). These data are recorded
automatically at the end of the SAVE file (see Section 8.4) upon completion of the
any TOUGH+ v1.5 run using a NAMELIST format. This is a very powerful
format that allows maximum clarity and flexibility, accepting free formats,
arbitrary ordering of variables, insertions of comments anywhere in the input
fields, and providing the option of ignoring any of the NAMELIST parameters by
omitting it or by not assigning a value to it. For more information, the reader is
directed to a textbook on FORTRAN 95/2003.
.
.
.
.
FJ000 0.00000000E+00 Aqu: P, X_m_A, X_i_A, T
3.2500495961160E+07 0.0000000000000E+00 3.0000000000000E-02 9.3974540000000E+00
FJ106 0.00000000E+00 Aqu: P, X_m_A, X_i_A, T
3.3467136230360E+07 2.9244292873310E-03 3.0000000000690E-02 1.2627840000000E+01
:::
&Data_For_Continuation_Run
timesteps_to_this_point = 6818,
NR_iterations_to_this_point = 71226,
number_of_detected_media = 8,
origin_of_time = 0.00000000000000E+00,
time_to_this_point = 4.69736464257198E+07,
accumulated_quantities = 7.35454091207164E+04, 1.08644220169487E+05,
1.03974248845536E+05, 6.43009774059305E+04,
0.00000000000000E+00, 0.00000000000000E+00,
/
<<<

Figure 8.1. An example of the NAMELIST-described termination data printed at the end of the
SAVE file from a TOUGH+ v1.5 simulation. These data can be read as part of the INCON data
block, or of the INCON external file. The names of the variables defined in the NAMELIST are self-
explanatory. For reference, this figure lists the conditions in the last two elements (FJ000 and
FJ106) of the grid in the TOUGH+ v1.5 simulation.

129
8.2. Data Block INDOM

This block introduces domain-specific initial conditions. These will supersede default

initial conditions specified in PARAM.4 (see Section 10), and can be overwritten by

element-specific initial conditions in data block INCON or data block EXT-INCON. The

option START is needed to use INDOM conditions.

Record INDOM.l

Format (A5, 2X, A3)

Rk_name, StateIndex

Rk_name
The name of a medium (corresponding to a system subdomain), as
specified in data block ROCKS.

StateIndex
The state identifier describing the conditions applying to the Rk_name
medium/subdomain.

Record INDOM.2

Format (6E20.13)
X(i), i = 1,NumCom+1

X(i) are the primary variables assigned to all grid blocks in the domain specified
in record INDOM.l. See Section 3.1 (Tables 3.1 and 3.2) for description of the
potential sets of primary variables.

Record INDOM.3

A blank record closes the INDOM data block.

Repeat records INDOM.l and INDOM.2 for as many domains as desired. The
ordering is arbitrary and need not be the same as in block ROCKS.

130
8.3. Data Block EXT-INCON

This block introduces extended capabilities for specifying initial conditions of subdomains

(i.e., groups of element). The user has several options to provide element information

(e.g., an element name or number list, location, element sequence, columnar structure,

etc.) that defines a subdomain to be initialized with the specified conditions. The first

entry is the total number of initialization entities entered within the block. Following that,

each of the various input entities may be entered in an arbitrary order.

Record EXT-INCON.0
This record includes the single integer variable Total_input_num that
describes the total number of input entities (representing initial conditions in
particular subdomains) that will be entered. This is read using a free format.

Record EXT-INCON.1
This record includes the single character variable INTYPE that is read using a free
format, and which provides a keyword defining the type of data describing the
subdomain to be initialized. INTYPE can assume one of the following values:
'GEOMETRY', 'LIST', 'SEQUENCE', or 'COLUMN'. Thus, INTYPE determines
what kind of variables/data will be read in record in EXT-INCON.2.

8.3.1. Data Block GEOMETRY

The data in this block are read when INTYPE = 'GEOMETRY', and define a set of

elements in a subdomain that is bounded within prescribed minimum and maximum

coordinates. All elements within this range will be assigned the initial conditions entered

in the record EXT-INCON.3.

Record EXT-INCON.2.0

For a Cartesian grid (coordinate_system = 'Cylindrical', see Section

5), the following real variables are read in EXT-INCON.2 using a free format:

131
 Xmin, Xmax, Ymin, Ymax, Zmin, Zmax

These parameters are defined as follows:

Xmin, Xmax
Real variables indicating the range of the subdomain to-be-initialized along
the x-axis of the Cartesian coordinate system.
Ymin, Ymax
Real variables indicating the range of the subdomain to-be-initialized along
the y-axis of the Cartesian coordinate system.
Zmin, Zmax
Real variables indicating the range of the subdomain to-be-initialized along
the z-axis of the Cartesian coordinate system.

For a cylindrical grid (coordinate_system = 'Cartesian', see Section

5), the following real variables are read in EXT-INCON.2 using a free format:

Rmin, Rmax, Zmin, Zmax

These parameters are defined as follows:

Rmin, Rmax
Real variables indicating the range of the subdomain to-be-initialized along
the r-axis of the cylindrical coordinate system.

Zmin, Zmax
Real variables indicating the range of the subdomain to-be-initialized along
the z-axis of the cylindrical coordinate system.

8.3.2. Data Block SEQUENCE

For INTYPE = 'SEQUENCE'

Record EXT-INCON.2.0

Format (*), i.e., free format

SequICFirstElemNum,
SequICLastElemNum,
SequICStride

These parameters are defined as follows:

132
 SequICFirstElemNum
Integer describing the global number of the first element in the sequence.

SequICLastElemNum
Integer describing the global number of the last element in the sequence.
SequICStride
Integer describing the stride in the numbering sequence
Thus, a sequence of elements is defined by the beginning and ending element
number, as well as by the stride (number of elements to skip between two
successive elements in the subdomain defined by the sequence). For example, a
sequence with SequICFirstElemNum = 10, SequICLastElemNum =
20, and SequICStride = 2 results in the a group (subdomain) of elements
with the following global element numbers: 10, 12, 14, 16, 18, 20. If
SequICLastElemNum < SequICFirstElemNum or if SequICStride
< 0, an error message is printed and the simulation is aborted.

All elements within this sequence will be assigned the initial conditions entered in
the record EXT-INCON.3.

8.3.3. Data Block LIST

For INTYPE = 'LIST'

Record EXT-INCON.2.0
Format (*), i.e., free format
ListICLength, N_per_row

These parameters are defined as follows:

ListICLength
Integer denoting the length of the list (i.e., the total number of element
numbers in the list) that is about to be read.

N_per_row
Integer defining the number of entries (= element numbers) per row in the
list that that is about to be read. The last row may have fewer than
N_per_row entries.

133
 Record EXT-INCON.2.1
Format (*), i.e., free format
ElemNum(i), i=1, ListICLength

NOTE: The number of entries per row of the ElemNum(i) input data is
N_per_row

The element numbers that are defined through their participation in a list are read
in rows (= records), each (except possibly the last one) containing N_per_row
entries. The last row/record may have fewer than N_per_row entries. For
example, if ListICLength = 8 and N_per_row = 3, then the element
numbers would be listed as follows:
ElemNum(1),ElemNum(2),ElemNum(3), ! 1st record
ElemNum(4),ElemNum(5),ElemNum(6), ! 2nd record
ElemNum(7),ElemNum(8) ! 3rd record
All elements in this list will be assigned the initial conditions entered in the
following record (EXT-INCON.3).

8.3.4. Remaining Data Blocks in EXT-INCON

For INTYPE = 'GEOMETRY', 'SEQUENCE', or 'LIST'

Record EXT-INCON.3

Format (*), i.e., free format

StateIndex, X0(i),i=1,NumCom+1

These parameters are defined as follows:

StateIndex
A character variable indicating the state index (see Section 3)
corresponding to the initial conditions that are to be assigned to the
subdomain defined by INTYPE.
X0(i),i=1,NumCom+1
A real array that includes the primary variables (corresponding to
StateInd) describing the conditions to which the respective subdomain
is initialized. See Section 3.1 for a thorough description of the potential
sets of primary variables.

Repeat records EXT-INCON.3, EXT-INCON.2 and EXT-INCON.3 for a total

of Total_input_num subdomain-based initializations.

134
8.3.5. Data Block COLUMN

This data block is somewhat different from the previous ones, and is very useful in

applying initial conditions in problems that involve initialization after achieving gravity

equilibration in a single column of the domain, which then serves as the reference column.

Thus, when INTYPE = 'COLUMN', the following records and data are read:

Record EXT-INCON.2.0

This record includes the single integer variable ColmICSize that is read using a
free format, and which describes the column length, i.e., the number of elements in
the column.

Record EXT-INCON.2.1

Format (*), i.e., free format

ColmICFirstElemNum,
ColmICNumElemInCol,
ColmICstride

These parameters are defined as follows:

ColmICFirstElemNum An integer variable indicating the global element

number of the first element in the columnar structure
to be initialized.

ColmICNumElemInCol An integer variable describing the total number of

elements to be initialized using the columnar
structure. Thus, the total number of columns to be
initialized using the data provided in this block is:
ColmICNumElemInCol/ColmICSize

ColmICstride An integer variable describing the stride in the

numbering sequence. This number is the difference
between the global numbers of two elements that
two successive locations in the same column.

For example, if ColmICSize=10, ColmICFirstElemNum=1,

ColmICNumElemInCol=80, and ColmICstride=1, then initialization using
the COLUMN data block will assign the initial conditions (obtained from the 10
elements of the reference column) to columns composed of elements with the

135
 following global numbers: Column #1, elements 1 to 10; Column #1, elements 11
to 20; … Column #8, elements 71 to 80. In this case, elements 1, 11, 21,…,71
have the same initial conditions (equal to those of the first entry in the reference
column). Similarly, elements 2, 12, 22, …, 72 have all the same initial conditions
(equal to those of the second entry in the reference column).

Conversely, if ColmICstride=1, then elements 1 to 8 have all the same initial

conditions (equal to those of the first entry in the reference column), 2 to 16 have
all the same initial conditions (equal to those of the second entry in the reference
column), etc.

Record EXT-INCON.3.x, x=1,…,ColmICSize

Format (*), i.e., free format

StateIndex, (X0(i),i=1,NumCom+1)

A total of ColmICSize records are read, each providing the state index and
primary variables of the elements of the reference column. The variables in these
records are as previously defined (see Section 8.3.4).

8.4. Implementing Initial Conditions

Flow systems are initialized by assigning a complete set of primary thermodynamic

variables to all grid blocks into which the flow domain is discretized. Various options are

available in a hierarchical system, as follows. During the initialization of a TOUGH+ run,

all grid blocks are first assigned to default thermodynamic conditions specified in data

block PARAM. The defaults can be overwritten for selected reservoir domains by assigning

domain-specific conditions in data blocks INDOM or EXT_INCON. These in turn may be

superseded by thermodynamic conditions assigned to individual grid blocks in data block

INCON. A disk file INCON written to the same specifications as data block INCON may

also be used.

136
 The possible sets of primary variables are discussed in Section 3.1 (Tables 3.1 and

3.2), with the actual primary variables depending on the fluid/solid phase composition.

During phase change primary variables will be automatically switched from one set to

another. In multiphase flow systems, therefore, different grid blocks will in general have

different sets of primary variables, and must be initialized accordingly.

For many applications, special initial conditions are needed, such as gravity-

capillary equilibrium, or steady state corresponding to certain mass and heat flows. This

can be realized by performing a series of TOUGH+ runs, in which thermodynamic

conditions obtained in one run, and written to disk file SAVE, are used as initial conditions

in a subsequent continuation run. For example, in a hydrate accumulation simulation, a

first run may be made to obtain hydrostatic pressure conditions. These may subsequently

be used as boundary conditions in a second run segment to simulate undisturbed natural

state conditions with through-flow of mass and heat. This could be followed by a third run

segment with fluid production and injection.

Restarting of a TOUGH+ run is accomplished by renaming the file SAVE

generated in a previous run as file INCON for initialization. Usually additional (often

minor) adjustments will be made for a restart. For example, different specifications for the

number of time steps and desired printout times may be made. Some editing of the MESH

file may be needed to make certain grid blocks inactive, so that previously calculated

pressures can serve as boundary conditions (see below). In a continuation run, simulation

time and time step counters may be continuously incremented, or they may be reset to

zero. For example, the latter option will be used when simulating production and injection

137
operations following preparation of a natural initial state, which may correspond to a large

simulation time.

As far as the internal workings of the code is concerned, there is no difference

between a fresh start of a simulation and a restart. The only feature that makes a

simulation a continuation run is that the INCON data were generated by a previous

TOUGH2 or TOUGH+ run, rather than having them explicitly provided by the user.

The file SAVE originating from TOUGH2 simulations [Pruess et al., 1991; 2012]

always ends with a data record with ‘+++’ in the first three columns, followed by one

record with restart information (time step and iteration counters, simulation time). The

file SAVE originating from a TOUGH+ simulation terminates with continuation data

written using the NAMELIST format shown in Figure 8.1. In either case, to reset all

counters and continuation data to zero when using SAVE as file INCON for another

TOUGH+ run, users can simply replace all records below the conditions of the last

element with a single blank record.

8.5. Implementing Boundary Conditions

8.5.1. General

Boundary conditions can be of two basic types. Dirichlet conditions prescribe

thermodynamic conditions, such as pressure, temperature, etc. on the boundary, while

Neumann conditions prescribe fluxes of mass or heat crossing boundary surfaces. A

special case of Neumann boundary conditions is no flux, which is the default in the

integral finite difference framework when no flow connections are specified across the

138
boundary. More general flux conditions are prescribed by introducing sinks or sources of

appropriate strength into the elements adjacent to the boundary.

In TOUGH2 [Pruess et al., 1999; 2012], Dirichlet conditions could be

implemented by assigning very large volumes (e.g., V = 1050 m3, as described by the

elem_vol variable in Section 7.2) to grid blocks adjacent to the boundary, so that their

thermodynamic conditions do not change at all from fluid or heat exchange with finite-

size blocks in the flow domain. In addition, a small value (such as 10-9 m) should be

specified for the nodal distance (ConxD1 or ConxD2, see Section 7.3) of such blocks, so

that boundary conditions are in fact maintained in close proximity to the surface where

they are desired, and not at some distance from it. It is possible to specify a nodal distance

that is outright zero; however, this may interfere with options for the computation of

interface mobilities that are intended for modeling fracture-matrix interactions. Therefore,

assigning zero nodal distances (ConxD1 or ConxD2) should be used with caution.

For time-independent Dirichlet boundary conditions, TOUGH2 [Pruess et al.,

1999; 2012] offered an alternative implementation, which provided savings in

computational work along with added user conveniences in running simulation problems.

This was accomplished by defining active and inactive elements. By convention,

elements encountered in data block ELEME (or in geometry files MESH or MINC) were

taken to be active until the first element entry with a zero or negative volume was

encountered. The first element with volume elem_vol ≤ 0.0E0, and all subsequent

elements, were by convention taken to be inactive. The easiest way to declare selected
€
grid blocks as inactive was to use a text editor to move them to the end of the ELEME data

block, and then insert a dummy grid block of zero volume in front of them.

139
 TOUGH+ v1.5 maintains these two older options available in TOUGH2, but

provides an additional one that is much simpler as it does not require any editing of the

element volume or physically moving any portions of the element list. In TOUGH+ v1.5,

elements can be designated as inactive if the parameter element_activity in record

ELEME.1 (see Section 7.2) is set to 'I' (when its conditions are time-invariant) or

'Vxx' (when its conditions and properties vary over time). Then, these elements, as well

as all elements designated as inactive by the older TOUGH2 options, are treated as

inactive.

For the inactive elements no mass or energy balance equations are set up, their

primary thermodynamic variables are not included in the list of unknowns, and their

thermodynamic conditions remain unchanged during the course of the simulation.

Inactive elements can appear in flow connections and initial condition specifications like

all other elements. The computational overhead of inactive elements is small because they

do not increase the number of equations to be solved in a flow problem.

8.5.2. Data Block BOUNDARIES

This data block provides time-variable boundary conditions that are applicable to

gridblocks with an activity indicator elem_activity = ‘Vxx’ (xx is a number, see

Section 7.2). When this data block is present in the input data file, the time-variable

boundary conditions are read in a tabular form.

The data in this record are read using a NAMELIST format, and may occupy one or

more lines (a choice left to the user). As already discussed (Section 8.1), NAMELIST-

based formats are a feature of FORTRAN 95/2003 and provide unique power and

140
flexibility, allowing (a) assignment of updated values to any subset of the parameters

included in the NAMELIST definition, (b) arbitrary order, (c) free formats of individual

parameter values, (d) inclusion of comments, etc. Future versions of TOUGH+ will

involve NAMELIST-based formats to read most input data.

The following records and data are read in the ‘BOUNDARIES’ data block:

Record BOUNDARIES.1

The namelist in this record is named Transient_Boundaries, has the

following general form

&Transient_Boundaries number_of_defined_boundaries = x,
num_table_defined_boundaries = x /

and includes the following variables:

number_of_defined_boundaries:
An integer variable describing the number of time-variable boundaries that
are to be described
num_table_defined_boundaries:
An integer variable describing the number of tables that are to be read,
which describe the time-variable behavior of these boundaries.

The variable number_of_defined_boundaries can be equal or smaller

than num_table_defined_boundaries. However, the current TOUGH+
v1.5 version can only accommodate tabular data in the BOUNDARIES.1 block,
requiring that:

number_of_defined_boundaries = num_table_defined_boundaries

Record BOUNDARIES.2
This record includes general data defining the boundary. A total of
num_table_defined_boundaries such records need to be provided. The
namelist in this record is named Transient_Boundary_Definition, has
the form

&Transient_Boundary_Definition boundary_number = x,
boundary_name = 'x',
num_PV_to_read = x,
PV_numbers = x,
data_form = 'x'/

141
 and includes the following parameters:

boundary_number:
Integer parameter describing the number of the boundary (more than one
boundaries can be defined).
boundary_name:
The name of of the boundary corresponding to boundary_number.
num_PV_to_read:
Integer parameter describing the number of the primary variables to be
used from among the data read from the table. More than one primary
variables can be used to define a boundary, and the maximum number is
num_PV_to_read = NumEqu.
PV_numbers:
The number of the primary variables (as listed in the order of the primary
variables in the EOS of the User’s Manuals of the individual TOUGH+
application options) corresponding to each column of the tabular data.
data_form:
This character parameter (5 characters long) describes the type of the data
defining the boundary. In the current version of TOUGH+, the only option
is data_form = 'Table'.

The structure of the Transient_Boundary_Definition namelist (and its

use as an input format in the data block BOUNDARIES) is best illustrated in the
example of Figure 8.2.

Record BOUNDARIES.3

This record includes specific data describing the tabular data of the conditions at
the boundary. A total of num_table_defined_boundaries such records
need to be provided. The namelist in this record is named
Tabular_Data_Definition, has the form

&Tabular_Data_Definition read_by_row_or_by_col = x,
number_of_rows = 'x ',
tot_number_of_columns = x,
RowColNum_to_PrimVarNum = x,…,x,
units_of_PrimVar = 'x',…,'x',
read_format = 'x'/

and includes the following parameters:

read_by_row_or_by_col:
Character parameter of length 3 describing if the tabular data are to be read
by row or by column. The following self-explanatory options are available:

142
 ='Row': The data are read by row. Each row includes the time and the
corresponding primary variable value(s); each column corresponds to a
single time.

='Col': The data are read by column. Each column lists the entire series
of time or of one or more primary variables.
number_of_rows:
Integer parameter defining the number of data points.
tot_number_of_columns:
Integer parameter describing the number of the number of columns to be
read, including both the time and the boundary primary variable columns
(must be > 1).
units_of_PrimVar:
A character array of length 3 and of dimension NumEqu+1 that describes
the units of time and of the primary variables in the table.
read_format:
This character parameter (up to 120 characters long) describes the format to
read the tabular data.

The structure and use of the namelists in the data block BOUNDARIES is best
illustrated by the example of Figure 8.2.

Record BOUNDARIES.4
This record includes a comment (a character variable of 200 characters) that
provides some information on the table to follow.

Record BOUNDARIES.5.1, 5.2, 5.3, …, number_of_rows

Each one of those records includes tot_number_of_columns individual data
points of the table that are read using the format specified by read_format.

The structure and use of the data block BOUNDARIES is best illustrated by the
example of Figure 8.2.

143
 BOUNDARIES
&Transient_Boundaries number_of_defined_boundaries = 1,
num_table_defined_boundaries = 1
/
&Transient_Boundary_Definition boundary_number = 1,
boundary_name = 'V01'’,
num_PV_to_read = 2,
PV_numbers = 1,3,
data_form = 'Table'
/
&Tabular_Data_Definition read_by_row_or_by_col = 'row',
number_of_rows = 3,
tot_number_of_columns = 4,
RowColNum_to_PrimVarNum = 3,-1,0,1,
units_of_Primvar(0) ='sec',
units_of_Primvar(1) = 'Pa',
units_of_Primvar(3) = 'C',
read_format = '*'
/
'The PV are in the sequence: Temperature, dummy, time, dummy'
12.5e0 -1222.34 0.0e0 9.80e6
14.0e0 -1.48e03 3.0e3 8.80e6
17.0e0 -1.77e03 9.0e3 8.00e6

Figure 8.2. An example of the namelist structure of the BOUNDARIES data block. The time-
variable data are provided and read in tabular form.

144
 9.0. Sources and Sinks

9.1. Data Block GENER

This block introduces sinks and/or sources to the system

Record GENER.1 (for 5-character element names)

Format (A5, A5, 4I5, 5X,A4, A1, 3E10.4 ,A4, 6x, 3(E10.4))

ElName5C, SS_name, NSEQ, NADD, NADS, LTAB,

SS_Type, ITAB, GX, EX, HX,
WellResponse, PresLimits,
RateStepChange, RateLimit

These parameters are defined as follows:

ElName5C:
The code name of the 5-character element containing the sink/source.

SS_name:
The name of the sink/source. The first three characters are arbitrary, the
last two characters must be numbers.

145
 NSEQ: The number of additional sinks/sources with the same
injection/production rate (not applicable for SS_Type = 'DELV').

NADD: The increment between the code numbers of two successive elements
with identical sink/source.

NADS: The increment between the code numbers of two successive

sinks/sources.

LTAB: The number of points in table of generation rate versus time. Set 0 or 1
for constant generation rate. For wells on deliverability, LTAB denotes
the number of open layers, to be specified only for the bottommost layer.

SS_Type:
The type of source or sink. This variable specifies different options for
fluid or heat production and injection. For example, different fluid
components may be injected, the nature of which depends on the EOS
module being used. Different options for considering wellbore flow
effects may also be specified. The following options are available:

='HEAT': Introduces a heat sink/source. This option is to be used for

injection only.
='COM1': Indicates mass component #1 (usually water). This option is
to be used for injection only.
='WATE': Indicates water injection.
='COM2': Indicates mass component #2. This option is to be used for
injection only.
='COM3': Indicates mass component #3. This option is to be used for
injection only.
='COMn': Indicates mass component #n. This option is to be used for
injection only.
='MASS': Specified a mass production rate, i.e., the mass rate of all the
fluids withdrawn from a system.

ITAB: Unless left blank, a table of specific enthalpies will be read (LTAB > 1
only).

GX: The constant generation rate. GX is positive for injection and negative for
production. GX describes a mass rate [kg/sec] for generation types COMl,
COM2, COM3, etc., and MASS; it describes an energy rate [W] for a HEAT
sink/source. For wells on deliverability, GX is the productivity index PI
[m3] – see Equation (9.2).

EX: The fixed specific enthalpy [J/kg] of the fluid for mass injection (GX>0).
For wells on deliverability against fixed bottomhole pressure, EX is the

146
 bottomhole pressure Pwb [Pa] at the center of the topmost producing layer
in which the well is open.

HG: The thickness of layer [m]. This is to be used only in cases of wells on
deliverability with specified bottomhole pressure.

WellResponse:
A character variable that describes how the source/sink is to be treated if
pressure limits (as described by the variable PresLimits) at the
corresponding cell are violated. The following options are available:

='STOP': The simulation is halted.

='ZERO': The source/sink rate GX is reset to zero, and the simulation
continues.
='ADJU': The source/sink rate GX is adjusted, and the simulation
continues. Note that a simulation involving the
WellResponse = 'ADJU' option may involve several
successive GX adjustments.

PresLimits:
A real variable describing the pressure limit [Pa] that acts as a trigger for
WellResponse to be enacted. For injection, pressure in the element
containing the source has to exceed PresLimits for changes to be
implemented. For production, pressure in the element containing the sink
has to fall below PresLimits for changes to be implemented.

RateStepChange:
A real variable that is used only when WellResponse = 'ADJU'.
By convention, RateStepChange is always a positive number. It
represents the absolute value of the fraction by which GX is to decrease
when the PresLimits criterion is violated. For obvious reasons,
0 ≤ RateStepChange ≤ 1.

RateLimit:
A real variable describing the minimum rate limit, described as the lowest
fraction of the original GX to which the rate is allowed to decline. The
simulation stops when RateLimit is reached.

Record GENER.1 (for 8-character element names)

Format (A8, A5, 12X, I5, 5X,A4, A1, 3E10.4 ,A4, 6x, 3(E10.4))

ElName8C, SS_name, LTAB,

SS_Type, ITAB, GX, EX, HX,
WellResponse, PresLimits,

147
 RateStepChange, RateLimit

Here ElName8C is the name of the 8-character element containing the

sink/source. All other variables and parameters are as in the case of the 5-
character elements.

Record GENER.l.l (Optional, LTAB > l only)

Format (4E14.7)
Fl(k), k=l, LTAB

F1(k): Generation times in the table of the time variable source/sink data.

Record GENER.1.2 (Optional, LTAB > l only)

Format (4E14.7)
F2(k), k=1, LTAB

F2(k): Generation rates in the table of the time variable source/sink data.

Record GENER.1.3 (Optional, LTAB > l and ITAB non-blank only)

Format (4E14.7)
F3(k), k=1, LTAB

F3(k): Specific enthalpy of produced or injected fluid in the table of the time
variable source/sink data.

Repeat records GENER.1, GENER.1.1, GENER.1.2, and GENER.1.3 for the

desired number of sinks/sources.

Record GENER.2

A blank record closes the GENER data block. Alternatively, generation information
may terminate on a record with ‘+++’ typed in the first three columns, followed by
data describing the numbers of the elements where the sources/sinks are located.

148
9.2. Discussion on sinks and sources
Sinks and sources are introduced through data block GENER in the input file. Several

options are available for specifying the production (q < 0) or injection (q > 0) of fluids

and heat. Any of the mass components may be injected in an element at a constant rate, or

at time-dependent rates that may be prescribed through user-defined tables. The user has

the option of specifying the specific enthalpy of the injected fluid as either a constant or

time dependent value. Heat sources/sinks (with no mass injection) may be either constant

or time-dependent.

Fluid production from an element may be handled by prescribing a constant or

time-dependent mass rate. In this case, the phase composition of the produced fluid may

be determined by the relative phase mobilities in the source element. Alternatively, the

produced phase composition may be specified to be the same as the phase composition in

the producing element. In either case, the mass fractions of the components in the

produced phases are determined by the corresponding component mass fractions in the

producing element. Different options are available for interpolating time-dependent rates

from user-supplied tabular data; these may be selected through parameter MOP(12) – see

Section 10.

149
 PAGE LEFT INTENTIONALLY BLANK

150
 10. Computational Parameters

In this section the data blocks (PARAM and SOLVR) that specify computational

parameters are described and followed by a discussion on linear equation solvers.

10.1. Data Block PARAM

This block describes an assortment of computation parameters.

Record PARAM.1

Format (2I2, 3I4, 24I1, 3E10.4, 2I5)

Max_NumNRIterations, OutputOption,
Max_NumTimeSteps, iCPU_MaxTime,
PRINT_frequency, (MOP(i),i=1,24),
BaseDiffusionCoef, DiffusionExpon,
DiffusionStrength,
SAVE_frequency, TimeSeries_frequency

These parameters are defined as follows:

151
 Max_NumNRIterations:
Integer specifying the maximum number of Newtonian iterations per time
step (default is 8)

OutputOption:
Integer defining an option that controls the amount of printout in the
standard TOUGH+ output (the default is 1). The following options are
available:
= 0,1: Print a selection of the most important variables.
= 2: In addition, print the mass and heat fluxes and the flow velocities.
= 3: In addition, print the primary variables and their changes.
If the above values for OutputOption are increased by 10, printout will
occur after each Newton-Raphson iteration (not just after convergence).

Max_NumTimeSteps:
An integer defining the maximum number of time steps allowed during the
simulation.
NOTE: If Max_NumTimeSteps < 0, then the maximum number of
timesteps becomes 1000*ABS(Max_NumTimeSteps)

iCPU_MaxTime:
An integer describing the maximum duration, in CPU seconds, of the
simulation (default is infinite).

PRINT_frequency:
An integer describing the printout frequency. This, printout will occur for
every multiple of PRINT_frequency steps (the default is 1).
NOTE: If PRINT_frequency < 0, then the printout frequency
becomes 1000*ABS(PRINT_frequency)

MOP(i), i=1,24:
An integer array that allows choice of various computational options.
These are described in detail below, and are documented in the printed
standard output from a TOUGH+ v1.5 run.

MOP(1):
If MOP(1) ≠ 0, a short printout for non-convergent iterations will be
generated at the end of each Newton-Raphson iteration.

MOP(2)to MOP(6):
If ≠ 0, these options generate additional printout in various subroutines at
the end of each Newton-raphson iteration. This feature is not needed in

152
 normal applications, but can be convenient in the development of new
capabilities as they can be used to identify coding errors. The amount of
printout increases with the value of MOP(i). The user is encouraged to
consult the source code listings for details. Below we list the subroutines
corresponding to MOP(2)to MOP(6):

MOP(2):
Simulation_Cycle (executive subroutine that advances time and
controls the flow of data during the simulation)

MOP(3):
JACOBIAN_SetUp (routine computing the flow and accumulation terms
in the mass and energy balance equations).

MOP(4):
SourceSink_Equation_Terms (subroutine determining the
contribution of sinks/sources to the mass and energy balance equations).

MOP(5):
Equation_Of_State (routine describing the equation of state of the
system under investigation and computing all thermophysical properties).

MOP(6):
Solve_Jacobian_Matrix_Equation (routine that solves the linear
equations of the Jacobian matrix).

MOP(7):
If MOP(7) ≠ 0, a printout of the input data is provided in the standard
output file.

MOP(8):
It determines how relative permeability and capillary pressure are estimated
in the presence of solid phases (see discussion of Section 2.12). The
following options are available:

=0: Based on the OPM model; capillary pressure scaling based on EPM #1.

=1: Based on EPM #1 model; capillary pressure scaling based on EPM #1.

=2: Based on EPM #2 model; capillary pressure scaling based on EPM #2.

=3: Based on EPM #1 model, no capillary pressure scaling.

=4: Based on EPM #2 model, no capillary pressure scaling.

=9: Based on OPM model, no capillary pressure scaling.

153
 MOP(9):
It determines the composition of produced fluid with the MASS option - see
discussion on the data block GENER in Section 9. The relative amounts of
phases are determined as follows:
= 0: according to relative mobilities in the source element.
= 1: the produced source fluid has the same phase composition as the
producing element.

MOP(10):
It controls the selection of the interpolation formula for the composite heat
conductivity as a function of the various phase saturations. The following
options are available

= 0: kθ = kθ d + SA ( kθ w −kθ d ) + φ SI kθ I

= 1: kθ = kθ d +SA ( kθ w −kθ d ) + φ SI kθ I

= 2: kθ = kθ d + φ ∑ SA kθβ
β =1,...,N β
β ≠G

= 3: kθ = kθ d + φ ∑ SA kθβ
β =1,...,N β

Options MOP(10)=0 and MOP(10)=1 are based on extensions of an

earlier model of Somerton et al. [2003; 2004] and are applicable to two-
phase systems, in which SA represents the saturation of the wetting phase
(not necessarily the aqueous phase). It is not known under what conditions
(if any) the linear model of Bejan [1984] (invoked for MOP(10)= 2 and
3, indicating ignoring and accounting for the gas contribution) is
applicable, but it is included for completeness. The option MOP(10)= 3
is discouraged because of (a) doubts about the validity of the Bejan [1984]
linear model, (b) the very demanding computations for the estimation of
the gas thermal conductivity from the real gas property package in
TOUGH+ v1.5 is demanding, and (c) the small overall contribution to the
composite thermal conductivity.

MOP(11):
It provides alternative options for the evaluation of mobility and
permeability at interfaces. These are:

= 0: The mobilities are upstream weighted according to the

W_upstream factor (see discussion in PARAM.3), and the
permeability is upstream-weighted.

154
 = 1: The mobilities are averaged between adjacent elements, and the
permeability is upstream-weighted.
= 2: The mobilities are upstream weighted, and the permeability is
harmonic-weighted.
= 3: The mobilities are averaged between adjacent elements, and the
permeability is harmonic-weighted.
= 4: The mobility and permeability are both harmonic weighted.
For multiphase flow simulations in which the upstream element is not
known a priori, MOP(11) = 0 or MOP(11) = 2 are the recommended
options. The user is strongly cautioned against using other options.

MOP(12):
It determines the interpolation procedure of the tabular data (involving
times, flow rates and enthalpies, see Section 8) in time-dependent sources
and sinks. The following options are available:
= 0: triple linear interpolation; tabular data are used to obtain
interpolated rates and enthalpies for the beginning and end of the
time step; the average of these values is then used.
= 1: step function option; rates and enthalpies are taken as averages of
the table values corresponding to the beginning and end of the time
step.
= 2: rigorous step rate capability for time dependent generation data.
A set of times ti and generation rates qi provided in data block GENER is
interpreted to mean that sink/source rates are piecewise constant and
change in discontinuous fashion at table points. Specifically, generation is
assumed to occur at constant rate qi during the time interval [ti, ti+1), and
changes to qi+1 at ti+1. The actual rate used during a time step that ends at
time t, with ti ≤ t ≤ ti+1, is automatically adjusted in such a way that total
cumulative exchanged mass at time t
t i−1
Q(t) = ∫ q dt = ∑ q (tj j+1 − t j ) + qi (t − ti )
0 j=1

is rigorously conserved. If tabular data for enthalpies are also provided, an

analogous adjustment is made to fluid enthalpy to preserve ∫ qh dt .

MOP(13):
Option used in processes involving mechanical dispersion.

MOP(14):
It specifies the handling of gas solubility in liquid phases according to one
of the following options:

155
 = 0: The gas solubility is computed using appropriate equations of
Henry’s dissolution parameters (T-dependent).
> 0: The gas solubility is computed using fugacities (T- and P-dependent)
For low P and T variations, it is possible to use the MOP(14)=0 option,
which leads to the use of a P- and T-invariant Henry’s constants for the
estimation of solubility. When MOP(14)=0, solubility is computed from
fast parametric equations that describe the effect of temperature. For
MOP(14)>0, gas solubility is computed from fugacities and activity
coefficients. This option is not activated for all TOUGH+ v1.5 application
options; in several application options, it is not possible to enable the
computationally intensive MOP(14)>0 option because it is excessive and
unnecessary.

MOP(15):
A flag indicating whether conductive heat exchange with impermeable
confining layers (see Section 7.4) is to be considered.
= 0: The heat exchange is not considered.
= 1: The heat exchange is activated (for grid blocks that have a non-zero
heat transfer area; see data block ELEME in Section 7.2).
MOP(16):
It provides automatic time step control. Time step size will be doubled if
convergence occurs within ITER ≤ MOP(16) Newton-Raphson iterations.
It is recommended to set MOP(16) in the range of 2 - 5. The default
value is 4.

MOP(17):
It specifies the handling of binary gas diffusivities according to one of the
following options:
= 0: The binary gas diffusivity is computed from the method of Fuller et
al. [1969].
= 7: The basic estimate of the binary gas diffusivity is computed from
the method of Fuller et al. [1969], and is then adjusted for high
pressures using the method of Riazi and Whitson [1993].

MOP(18):
A flag determining the method for estimating interface density.
= 0: Perform upstream weighting for interface density.
> 0: Use the average interface density between the two grid blocks.
However, when one of the two phase saturations is zero, upstream
weighting is to be performed.

156
MOP(19):
This is the parameter that controls the simulation output. When
MOP(19)<8, a standard ASCII file output (as described in general terms
in all the User’s Manuals of the individual TOUGH+ v1.5 application
options) is produced. Depending on the value of the OutputOption
parameter in block PARAM.1, this output can includes pressure,
temperature and saturation distribution of the various phases,
concentrations, thermophysical properties, and primary and secondary
variables.
When MOP(19)=8, an additional file containing the most important
properties are also printed in a format that conforms to the requirements of
the TecPlot package [TecPlot, 2003], and is suitable for most other plotting
and graphing packages. The name of this file is Plot_Data_Elem, and
it stores the element-specific properties and parameters for plotting and
graphing. For MOP(19)=9, the plotting files and a truncated standard
output file are produced (listing only mass balances at the prescribed
printout times).

MOP(20):
Flag determining whether the validity of the initial conditions is to be
checked. The options are:
< 9: The initial conditions are checked to ensure physically meaningful
and non-contradictory state indices and the corresponding primary variable
values (default).

= 9: No checking of initial conditions is performed.

This option may be useful in continuation runs involving large grids, in
which case checking of the initial conditions (as provided by the SAVE file)
is both time consuming (in terms of computation time) and generally
unnecessary. In general, the user is discouraged from bypassing the
checking process.

MOP(21):
A computational parameter to selects the linear equation solver (see
Section 10.2) from among the following options:
= 0: defaults to MOP(21) = 3.
= 1: LUBAND, banded direct solver using LU decomposition.
= 2: DSLUBC, bi-conjugate gradient solver with preconditioner.
= 3: DSLUCS (default), Lanczos-type preconditioned bi-conjugate
gradient solver with preconditioner.

157
 = 4: DSLUGM, generalized minimum residual preconditioned conjugate
gradient solver with preconditioner.
= 5: DLUSTB, stabilized bi-conjugate gradient solver with
preconditioner.
All conjugate gradient solvers use incomplete LU-factorization as a default
preconditioner. Other preconditioners may be chosen by means of the data
block SOLVR (see Section 10.2).

MOP(22), MOP(23), MOP(24):

The function of these variables are described in the individual User’s
Manuals of the TOUGH+ v1.5 application options (EOS) in which they are
active/enabled.
BaseDiffusionCoef:
The base gas diffusion coefficient [m2]

DiffusionExpon:
Parameter (exponent) describing the temperature dependence of gas phase
diffusion coefficient – see Equation (6.4).
DiffusionStrength:
Parameter (optional) describing the effective strength of enhanced vapor
diffusion; if set to a non-zero value, it will replace the parameter group
φτ0τ for vapor diffusion – see Equations (2.15) and (2.17), and Section
β

6.4.1.
SAVE_frequency:
Frequency of writing and saving the SAVE file. This feature avoids data
loss if the simulation is interrupted. A value between 100 and 500 is
recommended. When SAVE_frequency = 0, the SAVE file is written
and stored only once at the conclusion of the simulation.

TimeSeries_frequency:
Frequency of writing and saving the data in the various time-series output
files tracking subdomains, interfaces and/or source-sink (well) groups (see
Sections 10.3 to 10.5). This feature avoids the creation of very large time
series files in long simulations that involve many thousands of time steps.
Note that mass-balance related calculations related to some of the
parameters included in the time series output files are conducted at the
conclusion of each time step, but the results are printed only when specified
by this parameter. The default value of TimeSeries_frequency = 1.

158
Record PARAM.2
Format (4E10.4, A5, 5X, 3E10.4)

TimeOrigin, SimulationTimeEnd,
InitialTimeStep, MaxTimeStep, TrackElemName,
gravity, Dt_reducer, scale

These parameters are defined as follows:

TimeOrigin:
A real variable indicating the origin of time (starting time) in the simulation
[sec]. The default is TimeOrigin = 0.0E0.
SimulationTimeEnd:
A real variable indicating the time [sec] at which simulation should stop.
The default is infinite.
InitialTimeStep:
A real variable specifying the initial time step size [sec]. If
InitialTimeStep<0, then the program proceeds to read NumDts =
INT(ABS(InitialTimeStep)) records with time step information.

MaxTimeStep:
A real variable defining the upper limit for time step size [sec]. The default
is infinite.

TrackElemName:
A character variable providing the name of an element, the behavior of
which is to be tracked over time by printing a short printout of the
evolution of its key conditions and properties after each time step.

Gravity:
A real variable specifying the magnitude [m/sec2] of the gravitational
acceleration vector. Blank or zero gives "no gravity" calculation.

Dt_reducer:
A real variable defining the factor by which time step is reduced in case of
convergence failure or other problems. The default value is 4.

Scale:
The scale factor (a real variable) by which the size of the mesh is adjusted
(default = 1.0).

159
 Record PARAM.3
Format (7E10.4, 2X, A3, 5X, 3E10.4)
rel_convergence_crit, abs_convergence_crit,
U_p, W_upstream, W_NRIteration,
derivative_increment, W_implicitness,
DefaultStateIndex, P_overshoot, T_overshoot,
S_overshoot
These parameters are defined as follows:
rel_convergence_crit:
Convergence criterion for relative error (real variable, parameter ε1, see
Equation (3.8), default = 10-5).
abs_convergence_crit:
Convergence criterion for absolute error (real variable, parameter ε2, see
discussion of Equation (3.9), default = 1).

U_p: Not used in TOUGH+ v1.5; maintained only to ensure compatibility with
older TOUGH2 [Pruess et al., 1999] input files.
W_upstream:
The upstream weighting factor (real variable) for computing mobilities and
enthalpies at interfaces. The default W_upstream = 1.0 is strongly
recommended for multi-phase flows (0 ≤ W_upstream ≤ 1).
W_NRIteration:
A weighting factor 0 < W_NRIteration ≤ 1 (real variable) determining
the level of updating of the solutions based on the results of the
Newton/Raphson iteration. The default W_NRIteration = 1.0E0 is
recommended.
derivative_increment:
The increment factor (a real variable) for numerically computing
derivatives. The default value is derivative_increment = 10-m/2,
where m, evaluated internally, is the number of significant digits of the
floating point processor used. For 64-bit arithmetic,
derivative_increment ≈ 10-8.
W_implicitness:
A weighting factor 0 < W_implicitness ≤ 1 (a real variable) describing
the level of implicitness in the solutions. The default W_implicitness
= 1.0 is recommended.
DefaultStateIndex:
The default state identifier (a character variable) of the general initial
conditions that apply uniformly over the entire if not amended by the data
blocks/files describing initial conditions in the domain (see Section 8).

160
P_overshoot:
A real variable specifying the level of overshoot (defined as a fraction)
allowed in the computation of pressure if P is used as a criterion for
triggering phase and state changes. When P_overshoot = 0.0e0
(default), the P-triggered phase and state changes are at their most accurate
and sensitive (hair-trigger).
This variable is introduced to alleviate potential problems caused by
narrow oscillations about phase equilibrium lines that are possible under
certain conditions. In most cases, a value of 10-6 ≤ P_overshoot ≤ 10-4
is sufficient if a hair trigger causes problems. If P_overshoot < 0 or
P_overshoot ≥ 5.0x10-2, it is rest internally to its default value (=0).
T_overshoot:
A real variable specifying the level of overshoot (defined as a fraction)
allowed in the computation of temperature if T is used as a criterion for
triggering phase and state changes. The definitions, defaults, limits and
application are entirely analogous to those of P_overshoot.
S_overshoot:
A real variable specifying the level of overshoot (defined as a fraction)
allowed in the computation of saturations if thearese used criteria for
triggering phase and state changes. The definitions, defaults, limits and
application are entirely analogous to those of P_overshoot.

Record PARAM.4

This record introduces a set of primary variables that are used as default initial
conditions for all grid blocks that are not assigned by means of data blocks
INDOM, INCON or EXT-INCON. The format and date read here are:

Format (6E20.13)
default_initial_cond(i), i=1,NumCom+1

As is self evident, the real variables default_initial_cond(i) describe the

initial conditions of the state defined by StateIndex, as defined by the
corresponding primary variables. When more than six primary variables are
needed, more than one line (record) must be provided. See Section 3.1 for
description of potential sets of state indices and primary variables.

161
10.2. Modification of Computational Parameters
During the Course of a TOUGH+ Simulation
It is possible to modify the computational parameters described in Section 10.1 in the

course of a TOUGH+ v1.5 simulation without having to interrupt the execution. This

feature is particularly useful in simulations involving large grids and a large number of

timesteps (Max_NumTimeSteps) when the user observe solution convergence and time

advancement that can be improved by varying some of these computational parameters.

The process is controlled by the parameter SAVE_frequency (see Section 10.1,

record PARAM.1) that determines the frequency of updating the SAVE file. At the time of

updating the SAVE file, TOUGH+ also interrogates the directory of execution for the

presence of a file called Parameter_Update_File. If no such file exists, there is no

updating of the computational parameters.

If a file Parameter_Update_File exists, then it is opened and the following

data are read:

Record UPDATE.1

This record includes the single character variable UpdateHeader that is read
using a free format. The TOUGH+ computational parameters are updated only
when UpdateHeader = 'Update_Simulation_Parameters'.
Otherwise, the Parameter_Update_File is closed and the simulation
continues without any parameter updating.

Record UPDATE.2

This record includes a set of real variables (computational parameters) that are to
be updated. The data in this record are read using a NAMELIST format, and may
occupy one or more lines (a choice left to the user). As already discussed (Section
8.1), NAMELIST-based formats are a feature of FORTRAN 95/2003 and provide
unique power and flexibility, allowing (a) assignment of updated values to any
subset of the parameters included in the NAMELIST definition, (b) arbitrary order,
(c) free formats of individual parameter values, (d) inclusion of comments, etc.

162
Future versions of TOUGH+ will involve NAMELIST-based formats to read most
input data.

The namelist in this record is named Real_Parameters_To_Update, and

includes the following real parameters:

(1) SimulationTimeEnd
(2) MaxTimeStep
(3) rel_convergence_crit
(4) abs_convergence_crit
(5) P_overshoot
(6) T_overshoot
(7) S_overshoot

The corresponding TOUGH+ computational parameters will be updated if values

are provided for any of these computational parameters. The structure of the
Real_Parameters_To_Update namelist (and its use as an input format) is
best illustrated in the example of Figure 10.1.

Record UPDATE.3

This record includes a set of integer computational parameters that are to be

updated. The data in this record are read using a NAMELIST format. This namelist
is named Integer_Parameters_To_Update, and includes the following
integer parameters:

(1) Max_NumTimeSteps
(2) Max_NumNRIterations
(3) MOP_16
(4) SAVE_frequency
(5) TimeSeries_frequency

The corresponding TOUGH+ integer parameters will be updated if values are

provided for any of these computational parameters. It is possible to stop a
simulation by providing a Max_NumTimeSteps value that is smaller than its
current value in the code. Then, the simulation will be halted upon reading the
smaller Max_NumTimeSteps value while at the same time preserving the data in
the SAVE file, which would be lost if the execution is interrupted.

The structure of the Integer_Parameters_To_Update namelist (and its

use as an input format) is best illustrated in the example of Figure 10.1.

Upon reading the contents of the Parameter_Update_File and updating the

computational parameters, TOUGH+ v1.5 (a) prints a prominent message in the
standard output file that provides all the new parameter values, and (b) replaces the

163
 UpdateHeader = 'Update_Simulation_Parameters' value with the
value UpdateHeader = '==> Completed Update # n', where n is the
number of the update. Because UpdateHeader has no longer the value that will
cause TOUGH+ to read the subsequent data, this substitution prevents multiple
readings of the same Parameter_Update_File while keeping track of the
number of updates and preserving the evolution of the updated parameters in the
input file. Note that several updates are possible in the course of a long simulation.
To accomplish this, the data for the next update are simply added to the top of the
Parameter_Update_File file without erasing the updating history up to this
point. The process is clearly illustrated in the example of Figure 10.2.

'Update_Simulation_Parameters' ! UpdateHeader
&Real_Parameters_To_Update ! Namelist #1
SimulationTimeEnd = 5.0d6 ,
MaxTimeStep = 3600. ,
rel_convergence_crit = 1.5d-5 ,
P_overshoot = 1.0e-4 ,
S_overshoot = 1.0e-5
/ ! Not updated: abs_convergence_crit, T_overshoot
&Integer_Parameters_To_Update ! Namelist #2
Max_NumTimeSteps = 500 ,
Max_NumNRIterations = 10000 ,
MOP_16 = 4 ,
SAVE_frequency = 100 ,
TimeSeries_frequency = 5
/

Figure 10.1. An example of a Parameter_Update_File for parameter updating in the course

of a TOUGH+ v1.5 simulation. Within the namelists (Real_Parameters_To_Update and
Integer_Parameters_To_Update), parameters can be entered in any order, data are read
using any kind of appropriate format, only the needed parameters are included, and comments
can be added.

164
 Update_Simulation_Parameters
&Real_Parameters_To_Update
SimulationTimeEnd = 3.0d7
/
&Integer_Parameters_To_Update
Max_NumTimeSteps = 2000
/
==> Completed Update # 3 ! 3rd Update
&Real_Parameters_To_Update
SimulationTimeEnd = 2.0d7 ,
MaxTimeStep = 8.64e4 ,
rel_convergence_crit = 2.5d-5 ,
/
&Integer_Parameters_To_Update
Max_NumTimeSteps = 1500 ,
Max_NumNRIterations = 20000
/
==> Completed Update # 2 ! 2nd Update
&Real_Parameters_To_Update
SimulationTimeEnd = 8.0d6 ,
MaxTimeStep = 7200. ,
rel_convergence_crit = 2.0d-5 ,
P_overshoot = 1.0e-5
/
&Integer_Parameters_To_Update
Max_NumTimeSteps = 1000 ,
MOP_16 = 5 ,
SAVE_frequency = 200 ,
TimeSeries_frequency = 5
/
==> Completed Update # 1 ! 1st Update
&Real_Parameters_To_Update
SimulationTimeEnd = 5.0d6 ,
MaxTimeStep = 3600. ,
rel_convergence_crit = 1.5d-5 ,
P_overshoot = 1.0e-4 ,
S_overshoot = 1.0e-5
/
&Integer_Parameters_To_Update
Max_NumTimeSteps = 500 ,
Max_NumNRIterations = 10000 ,
MOP_16 = 4 ,
SAVE_frequency = 100 ,
TimeSeries_frequency = 5
/

Figure 10.2. An example of a Parameter_Update_File indicating three completed parameter

updates, in addition to another one (at the top of the file) that has not yet been executed.

165
10.3. Data Block SOLVR

This (optional) block specifies parameters used by linear equation solvers.

Record SOLVR.1

Format(I1,2X,A2,3X,A2,2E10.4)
MatrixSolver, Z_preprocessing,
O_preprocessing, Max_CGIterationRatio,
CG_convergence_crit
These parameters are defined as follows:

MatrixSolver:
This integer variable selects the linear equation solver from among the
following options: = 1: LUBAND
= 2: DSLUBC
= 3: DSLUCS
= 4: DSLUGM
= 5: DLUSTB

Z_preprocessing:
A character variable that determines the type of
Z-preconditioning [Moridis and Pruess, 1998]. Regardless of user
specifications, Z-preprocessing will only be performed when iterative
solvers are used (2 ≤ MatrixSolver ≤ 5), and if there are zeros on the
main diagonal of the Jacobian matrix. The following options are available:
='Z0': No Z-preprocessing; default for NumEqu=1 and
MatrixSolver=1
='Z1': Replace zeros on the main diagonal by a small near-zero constant
(1.0E-25; default for NumEqu>1 and for
1<MatrixSolver≤5
='Z2': Make linear combinations of equations for each grid block to
produce non-zero main diagonal entries
='Z3': Normalize equations, followed by Z2
='Z4': Same as in O_preprocessing='O4'

O_preprocessing:
A character variable that determines the type of O-preconditioning
[Moridis and Pruess, 1998]. It can take the following possible values:

166
 ='O0': No O-preprocessing; default for NumEqu=1 and
MatrixSolver=1

='O1': Elimination of lower half of the main-diagonal submatrix with

center pivoting

='O2': O1 + Elimination of upper half of the main-diagonal submatrix

with center pivoting

='O3': O2 + Normalization; results in unit main-diagonal submatrices

='O4': Pre-processing which results in unit main-diagonal submatrices

without center pivoting

Max_CGIterationRatio:
An integer variable that specifies the maximum number of CG iterations as
a fraction of the total number of equations
(0 < Max_CGIterationRatio ≤ 1; default is
Max_CGIterationRatio = 0.1)

CG_convergence_crit:
A real variable that specifies the convergence criterion for the CG iterations
(1.0E-12≤CG_convergence_crit≤1.0E-6; the default
CG_convergence_crit=1.0E-6)

The solver DLUSTB implements the BiCGSTAB(m) algorithm [Sleijpen and

Fokkema, 1993], an extension of the BiCGSTAB algorithm of van der Vorst (1992).

DLUSTB provides improved convergence behavior when iterations are started close to the

solution, i.e., near steady state. The preconditoning algorithms can cope with difficult

problems in which many of the Jacobian matrix elements on the main diagonal are zero.

An example was given in Pruess et al. [1999; 2012] in “two-waters” problems in which

typically 2/3 of the elements in the main diagonal are zero. Our tests show that this type of

problem can be solved by means of Z2 or Z3 preconditioning [Moridis and Pruess, 1998].

167
10.4. Discussion on Linear Equation Solvers
The computational work to be performed in the course of a TOUGH+ v1.5 simulation

includes evaluation of thermophysical properties for all grid blocks, assembly of the

vector of residuals and the Jacobian matrix, and solution of the linear equation system for

each Newton-Raphson iteration step. Except for small problems with just a few grid

blocks, the most computationally intensive of these different tasks is the solution of the

linear equation system. TOUGH+ v1.5 offers a choice of direct or iterative methods for

linear equation solution; technical details of the methods and their performance can be

found in Moridis and Pruess [1998].

The most reliable linear equation solvers are based on direct methods, while the

performance of iterative techniques tends to be problem-specific and lacks the

predictability of direct solvers. The robustness of direct solvers comes at the expense of

large storage requirements and execution times that typically increase with problem size N

(= number of equations solved) proportional to N3. In contrast, iterative solvers have

much lower memory requirements, and their computational work will increase much less

rapidly with problem size, approximately proportional to Nω, with ω ≈ 1.4 - 1.6 [Moridis

and Pruess, 1995]. For large problems and especially 3-D problems with several thousand

grid blocks or more, iterative conjugate gradient (CG) type solvers are therefore the

method of choice.

The default linear equation solution technique in TOUGH+ uses DSLUCS, a

Lanczos-type bi-conjugate gradient solver, with incomplete LU-factorization as

preconditioner. Users need to beware that iterative methods can be quite “fickle” and may

fail for matrices with special features, such as many zeros on the main diagonal, large

168
numerical range of matrix elements, and nearly linearly dependent rows or columns.

Depending on features of the problem at hand, appropriate matrix preconditioning may be

essential to achieve convergence. Poor accuracy of the linear equation solution will lead to

deteriorated convergence rates for the Newtonian iteration, and an increase in the number

of iterations for a given time step. In severe cases, time steps may fail with residuals either

stagnating or wildly fluctuating. Information on the convergence of the linear equation

solution is written to disk file LINEQ, which may be examined with any text editor. Users

experiencing difficulties with the default settings are encouraged to experiment with the

various solvers and preconditioners included in the TOUGH+ package.

169
 PAGE LEFT INTENTIONALLY BLANK

170
 11. Output Specifications

In this section, the various primary and secondary variables that may be provided as

outputs from TOUGH+ v1.5 are discussed. In addition, data blocks are described for

specifying output data, such as the times at which data are printed (data block TIMES),

and for defining subdomains, interfaces and groups of sinks and sources (optional data

blocks SUBDOMAINS, INTERFACES and SS_GROUPS, respectively) where important

variables are to be monitored and tracked by printing time-series data in external output

files. With the exception of data block TIMES that was available in the TOUGH2 family

of codes [Pruess et al., 1991; 2012], these are new TOUGH+ v1.5 output capabilities

unavailable to any earlier version of the code. This section also includes a few comments

on error messages and warnings in the TOUGH+ v1.5 family of codes.

171
11.1 Output of Primary and Secondary Variables
The TOUGH+ v1.5 code can provide the following output:

1. The pressure, temperatures, saturations, and equilibrium pressure

distributions.
2. The mass fractions of the various components in the various phases.
3. Flows and velocities of the phases across the gridblock interfaces
(connections) of the domain.
4. The primary variables and their changes in the elements of the domain.
5. Capillary pressures and relative permeabilities.
6. Densities, viscosities (when mobile), and enthalpies of the various phases.
7. Dissociation reaction rates and the corresponding heat of dissociation.
8. Volume and mass balances of the phases and components in the domain.
9. Production rates and production composition at wells.
10. Time series of the evolution of the most important variables at user-
specified elements, connections and sources/sinks.

Of those possible outputs, (1), (8), (9), and (10) are always printed in the standard

TOUGH+ v1.5 output. The amount of the output is controlled by the parameter

OutputOption in the data block PARAM. In keeping with the TOUGH2 convention, if

the OutputOption values are increased by 10, printouts will occur after each iteration

(not just after convergence). The specific outputs are discussed in detail in the User’s

Manuals of the various TOUGH+ application options.

11.2. Data Block TIMES

This optional block permits the user to obtain printout at specified times. This printout will

occur in addition to printout specified in record PARAM.1.

Record TIMES.1

Format (2I5, 2E10.4)

NumPrintTimes, Max_NumPrintTimes,
TimeStepMax_PostPrint, PrintTimeIncrement

172
 These parameters are defined as follows:

NumPrintTimes:
The number of times provided on records TIMES.2, TIMES.3, etc., (see
below; restriction: NumPrintTimes ≤ 100).

Max_NumPrintTimes:
The total number of times at which an output is desired
(NumPrintTimes ≤ Max_NumPrintTimes ≤ 100; default is
NumPrintTimes = Max_NumPrintTimes).

TimeStepMax_PostPrint:
The maximum time step size after any of the prescribed times have been
reached (default is infinite).

PrintTimeIncrement :
Time increment for times with indices NumPrintTimes,
NumPrintTimes+1, ..., Max_NumPrintTimes.

Record TIMES.2, TIMES.3, etc.

Format (8E10.4)
(PrintTime(i), i = l, NumPrintTimes)

PrintTime(i):
A list of times (in ascending order) at which printout is desired. Note that,
if any PrintTime(i+1)< PrintTime(i), the resulting negative
timestep forces the simulation to stop.

11.3. Data Block SUBDOMAINS

This optional data block is a new TOUGH+ v1.5 feature that is unavailable to any earlier

version of the code. It allows monitoring of the evolution of the pore volume-averaged

and properties and conditions in a set of subdomains (subsets) of the global grid (domain).

The time series results are written in a number of separate output files that are equal to the

number of the subdomains (a file for each subdomain), and are named according to the

convention SubdomName_Time_Series, where 'SubdomName' is the name of the

173
subdomain under observation. If the keyword 'SUBDOMAINS' is present in the input

file, then the following data are read using NAMELIST formats:

Record SUBDOMAIN.1
This record includes general data describing the number of subdomains to be
monitored. The namelist in this record is named Subdomain_General_Info
and has the general form:
&Subdomain_General_Info number_of_subdomains = x /,

i.e., it contains only the following single parameter:

number_of_subdomains:
An integer parameter describing the number of subdomains on which time
series data are to be obtained. There is no upper limit in the number of
subdomains that can be defined.

NOTE: The number_of_subdomains parameter must be set to a

positive value for the simulation to continue.

The structure of the Subdomain_General_Info namelist and its use as the

input format in the data block SUBDOMAINS is best illustrated in the examples of
Figure 11.1.

Record SUBDOMAIN.2

The namelist in this record provides general information on each of the

subdomains. A total of number_of_subdomains such records/namelists must
be provided. This namelist is named Individual_Subdomain_Specifics,
has the general form
&Individual_Subdomain_Specifics subdomain_name = x,
number_of_regions = x /

and includes the following parameters:

subdomain_name:
The subdomain name is described by this character variable of a maximum
of 8 characters. Upon reading the variable, the TOUGH+ code creates an
output file named SubdomName_Time_Series, where
'SubdomName' is the value of subdomain_name. If this variable is
omitted in the input, then the TOUGH+ code assigns internally the name
SubdomName = 'SubDomNN', where 'NN' is the 2-digit subdomain
number in the subdomain definition sequence.

174
number_of_regions:
An integer parameter specifying the number of regions of which the
subdomain is composed. NOTE: The number_of_regions parameter
must be set to a positive value for the simulation to continue.
The structure of the Individual_Subdomain_Specifics namelist and its
use as an input format in the data block SUBDOMAINS is best illustrated in the
examples of Figure 11.1.

Record SUBDOMAIN.2.1

The namelist in this record is named Region_Specifics. It provides

information that allows the definition of a region within a subdomain and
identification of the elements belonging to it. A total of number_of_regions
such records must by provided in order to fully define the subdomain. In its most
complete form it has the following structure:
&Region_Specifics definition_mode = 'x ',
number_of_elements = x,
first_element_number = x,
first_element_name = 'xxxxx',
element_sequence_stride = x,
format_to_read_data = 'x ',
region_shape = 'x ',
Xmin = x.xEx,
Xmax = x.xEx,
Ymin = x.xEx,
Ymax = x.xEx,
Zmin = x.xEx,
Zmax = x.xEx,
Rmin = x.xEx,
Rmax = x.xEx,
top_cylinder_center_xyz = x.xEx,
bot_cylinder_center_xyz = x.xEx,
sphere_center_xyz = x.xEx
/

The great advantage of the use of namelists is that only their necessary components
may be used and be assigned values; the rest can be omitted or commented-out.
Additionally, the namelist components can be listed in any order. The parameters
in the Region_Specifics namelist are defined as follows:
definition_mode:
This character variable (up to 10 characters in length) describes the method
by which the region is defined. It can take the following possible values:
='Geometry': The region is defined by its geometric boundaries
='Sequence': The region is defined by a sequence of element numbers

175
 ='NumberList': The region is defined by a list of element numbers
='NameList': The region is defined by a list of element names

number_of_elements:
This integer parameter is needed as an input when definition_mode
≠'Geometry' and describes the number of elements in the region.

format_to_read_data:
This character variable (up to 50 characters long) is needed as an input
when definition_mode ='NumberList' or 'NameList', and
describes the format to read the list of number_of_elements element
numbers or names. The number_of_elements elements to be read
according to the format format_to_read_data are listed
immediately after the end of the namelist (see examples of Figure 11.1).

NOTE: no other parameters of those in the namelist are needed if

definition_mode ='NumberList' or 'NumberList'.

first_element_number, first_element_name,
element_sequence_stride:
These integer parameters are needed as inputs when definition_mode
='Sequence'. Using the provided first_element_number or
determining it from the first_element_name, a total of
number_of_elements element numbers are determining using the
element_sequence_stride.

NOTE: no other parameters of those in the namelist are needed if

definition_mode ='Sequence'.

region_shape:
This character variable (up to 9 characters in length) is needed as an input
when definition_mode ='Geometry' and describes the shape of
the region that is about to be defined. The following self-explanatory
options are available:

='Rectangle': This option can only be used in Cartesian grids, i.e., if

coordinate_system = 'Cartesian' (see Section 5.1).

='Cylinder': This option can only be used for either Cartesian or

cylindrical grids.

='Sphere': This option can only be used for either Cartesian or

cylindrical grids.

176
If region_shape = 'Rectangle' and coordinate_system =
'Cartesian', then the following parameters must be included in the
namelist:

Xmin, Xmax, Ymin, Ymax, Zmin, Zmax

These real parameters indicate the range (minimum and maximum) of the
region along the x-, y- and z-axis of the Cartesian coordinate system,
respectively.

If region_shape = 'Cylinder' and coordinate_system =

'Cartesian', then the following parameters must be included in the
namelist:

Rmin, Rmax, top_cylinder_center_xyz,

bot_cylinder_center_xyz

The real parameters Rmin and Rmax are the minimum and maximum
radii of the cylindrical region. The 1D arrays
top_cylinder_center_xyz and bot_cylinder_center_xyz
are of size 3. These are the (x,y,z) coordinates of the centers of the top and
bottom circular surfaces of the cylinder, respectively.

If region_shape = 'Cylinder' and coordinate_system =

'Cylindrical', then the following parameters must be included in the
namelist:

Rmin, Rmax, Zmin, Zmax

These real parameters indicate the range (minimum and maximum) of the
region along the r- and z-axis of the Cartesiancylindrical coordinate system,
respectively.
If region_shape = 'Sphere' and coordinate_system =
'Cartesian' or coordinate_system = 'Cylindrical', then
the following parameters must be included in the namelist:
Rmin, Rmax, sphere_center_xyz
The real parameters Rmin and Rmax are the minimum and maximum
radius of the sphere. The (x,y,z) coordinates of the center of the sphere are
stored in the elements of the real array top_cylinder_center_xyz
of size 3. Of those three values, only the 3rd (corresponding to the z-
coordinate of the center of the sphere) one is used when
coordinate_system = 'Cylindrical'.
NOTE: no other parameters of those in the namelist are needed if
definition_mode ='Geometry'.

177
 The use of these inputs in the SUBDOMAINS datablock is best illustrated in the
examples of Figure 11.1.

NOTE #1: It is possible to combine regions of very different geometries to create

subdomains that are very irregular in shape. Each one of these regions can be
defined independently, i.e., by using different definition_mode values.
NOTE #2: All the various parameter combinations and permutations in the user-
supplied namelist Region_Specifics must be defined for the simulation to
continue.
NOTE #3: All the regions within the subdomain must be defined. A total of
number_of_regions records (one for each region of the subdomain) must be
provided in SUBDOMAIN.2.1. Thus, if we define two subdomains, the first with
3 regions and the second with 4, then we need to provide the following data:
For the 1st Subdomain: One SUBDOMAIN.2 record, and 3 SUBDOMAIN.2.1
records. For the 2nd Subdomain: One SUBDOMAIN.2 record, and 4
SUBDOMAIN.2.1 records.

11.4. Data Block INTERFACES

This optional data block is a new TOUGH+ v1.5 feature that is unavailable to any earlier

version of the code. It allows monitoring the flow – instantaneous and cumulative –

through user-defined interfaces (composed of smalled individual surfaces). The time

series results are written in a number of separate output files that are equal to the number

of the interfaces (a file for each interface), and are named according to the convention

InterfName_Time_Series, where 'InterfName' is the name of the interface

under observation. The InterfName_Time_Series files are important in continuatin

runs: information from the last TOUGH+ simulation is gleaned from the older files and is

used to seamlessly continue the computations of the cumulative mass flows through the

interfaces.

If the keyword 'INTERFACES' is present in the input file, then the following

data are read using NAMELIST formats:

178
SUBDOMAINS
&Subdomain_General_Info number_of_subdomains = 1 /
&Individual_Subdomain_Specifics subdomain_name = 'Zone1',
number_of_regions = 1 /
&Region_Specifics definition_mode = 'Geometry',
region_shape = 'Rectangle',
Xmin = 2.0e-1, Xmax = 4.0e-1,
Ymin = -1.0e8, Ymax = 1.0e8,
Zmin = -1.0e8, Zmax = 1.0e8,
! Rmin, Rmax, ! Not used – commented out
! top_cylinder_center_xyz, ! Not used – commented out
! bot_cylinder_center_xyz, ! Not used – commented out
/

SUBDOMAINS
&Subdomain_General_Info number_of_subdomains = 1 /
&Individual_Subdomain_Specifics subdomain_name = 'Zone8',
number_of_regions = 2
/
&Region_Specifics definition_mode = 'Sequence',
number_of_elements = 5,
first_element_number = 8,
element_sequence_stride = 1
/
&Region_Specifics definition_mode = 'Sequence',
number_of_elements = 5,
first_element_name = 'A0010',
element_sequence_stride = 1
/

&Subdomain_General_Info number_of_subdomains = 1 /
&Individual_Subdomain_Specifics subdomain_name = 'Zone5',
number_of_regions = 3 /
&Region_Specifics definition_mode = 'NumberList',
number_of_elements = 5,
format_to_read_data = '*',
/
8 9 10 11 12
&Region_Specifics definition_mode = 'NameList',
number_of_elements = 5,
format_to_read_data = '*',
/
'A0012' 'A0013' 'A0014' 'A0015' 'A0016'
&Region_Specifics definition_mode = 'Geometry',
region_shape = 'Cylinder'
Rmin = 2.0e-1, Rmax = 1.0e00,
! Ymin = -1.0e8, Ymax = 1.0e8, ! Not used – commented out
Zmin = -1.0e8, Zmax = 1.0e8,
! top_cylinder_center_xyz, ! Not used – commented out
! bot_cylinder_center_xyz ! Not used – commented out
/

Figure 11.1. Examples of the SUBDOMAINS data block for tracking the evolution of volume-
averaged properties and conditions in specified subdomains. This data block uses namelist-based
formats for data inputs.

179
 Record INTERFACE.1

This record includes general data describing the number of interfaces to be

monitored. The namelist in this record is named Interface_General_Info
and has the general form:
&Interface_General_Info number_of_interfaces = x /,

i.e., it contains only the following single parameter:

number_of_interfaces:
An integer parameter describing the number of interfaces at which time
series data are to be obtained. There is no upper limit in the number of
interfaces that can be defined.

NOTE: The number_of_interfaces parameter must be set to a

positive value for the simulation to continue.

The structure of the Interface_General_Info namelist and its use as the

input format in the data block INTERFACES is best illustrated in the examples of
Figure 11.2.

Record INTERFACE.2

The namelist in this record provides general information on each of the interfaces.
A total of number_of_interfaces such records/namelists must be provided.
This namelist is named Individual_Interface_Specifics, has the
general form

&Individual_Interface_Specifics interface_name ='x',

number_of_surfaces = x,
sign_of_flow_direction ='x'/

and includes the following parameters:

interface_name:
The interface name is described by this character variable of a maximum of
8 characters. Upon reading the variable, the TOUGH+ v1.5 code creates an
output file named InterfName_Time_Series, in which
'InterfName' is the value of interface_name. If this variable is
omitted in the input, then the TOUGH+ code assigns internally the name
InterfName = 'IntrFcNN', where 'NN' is the 2-digit interface
number in the interface definition sequence.

180
number_of_surfaces:
An integer parameter specifying the number of individual surfaces of which
the interface is composed. NOTE: The number_of_surfaces
parameter must be set to a positive value for the simulation to continue.

sign_of_flow_direction:
A character parameter of length 3 that defines how the sign of the flow
(positive or negative) is to be treated in reporting the information. It can
take the following possible values:

='ABS': The absolute value of the flows at the surface connections is

used. This is necessary in cases where the arbitrariness in the
order of the elements in a connection (a TOUGH+ v1.5
feature) may produce erroneous results if the arithmetic sum of
the flows through individual connections is used. For example,
in computing flow toward a well in a 3D Cartesian grid, the
direction of flow is well known but the signs of the individual
flows through each connection in the surface may differ
because of the element orientation. In such a case, the 'ABS'
value is known to accurately describe flows.

='DIR': The internal signs of the individual flows at the surface

connections are maintained in the flow summations. This
option is used when the order of elements in the definition of
the connections is known to be monotonic (in terms of
direction), i.e., when the grid numbering is predictably
consistent and the flow through a surface is known to be
variable in direction (e.g., the net recharge of an aquifer from
an aquitard with an operating injection well – water escapes
into int overlying aquifer near the injection well, but recharge
from the aquitard occurs away from the well).

The structure of the Individual_Interface_Specifics namelist and its

use as an input format in the data block INTERFACES is best illustrated in the
examples of Figure 11.2.

Record INTERFACE.2.1

The namelist in this record is named Surface_Specifics. It provides

information that allows the definition of a surface within an interface and the
identification of the connections belonging to it. A total of
number_of_surfaces such records must by provided in order to fully define
the interface. In its most complete form it has the following structure:

181
 &Surface_Specifics definition_mode = 'x',
number_of_connections = x,
format_to_read_data = 'x',
surface_shape = 'x',
Xmin = x.xEx,
Xmax = x.xEx,
Ymin = x.xEx,
Ymax = x.xEx,
Zmin = x.xEx,
Zmax = x.xEx,
surface_location = x.xEx,
surface_coord_number = x.xEx,
cylinder_radius = x.xEx,
top_cylinder_center_xyz = x.xEx,
bot_cylinder_center_xyz = x.xEx,
inner_circle_radius = x.xEx
outer_circle_radius = x.xEx
sphere_center_xyz = x.xEx
sphere_radius = x.xEx
/

The great advantage of the use of namelists is that only their necessary components
may be used and assigned values; the rest can be completely omitted or
commented-out. Additionally, the components of the namelist can be listed in any
order. The parameters in the Surface_Specifics namelist are defined as
follows:

definition_mode:
This character variable (up to 10 characters in length) describes the method
by which the region is defined. It can take the following possible values:
='Geometry': The surface is defined by its geometric attributes
='NumberList': The region is defined by a list of connection numbers
='NameList': The region is defined by a list of connection names

number_of_connections:
This integer parameter is needed as an input when definition_mode
≠'Geometry' and describes the number of connections in the surface.
format_to_read_data:
This character variable (up to 50 characters long) is needed as an input
when definition_mode ='NumberList' or 'NameList', and
describes the format to read the list of number_of_connections
connection numbers or names. The number_of_connections
elements to be read according to the format format_to_read_data
are listed immediately after the end of the namelist (see examples of
Figure 11.2).

182
 NOTE: No other parameters of those in the namelist are needed if
definition_mode ='NumberList' or 'NumberList'.

surface_shape:
This character variable (up to 9 characters in length) is needed as an input
when definition_mode ='Geometry' and describes the shape of
the surface that is about to be defined. The following self-explanatory
options are available:

='Rectangle': This option can only be used in Cartesian grids, i.e., if

coordinate_system = 'Cartesian'(see Section 5.1).

='Cylinder': This option can only be used for either Cartesian or

cylindrical grids.

='Circle': This option can only be used for either Cartesian or

cylindrical grids. It is used to compute flows through a circular surface,
e.g., at the top a bottom permeable boundaries of an aquifer with a well at
its center.

='Sphere': This option can only be used for either Cartesian or

cylindrical grids.

If surface_shape = 'Rectangle' and 'coordinate_system

= 'Cartesian', then the following parameters must be included in the
namelist:

Xmin, Xmax, Ymin, Ymax, Zmin, Zmax

The first four real parameters in the list indicate the range (minimum and
maximum) of the surface along the x-, y- and z-axis of the Cartesian
coordinate system, respectively. Obviously, the max and min values along
one (only) direction have to be identical, thus defining the orientation of the
planar surface. For example, a planar surface perpendicular to the z-axis is
defined when Zmin = -20m and Zmax = -20m. If the max and min
coordinate values match in more than one direction, or if there is no match
in any direction, an error message is printed and the simulation is aborted
because the orientation of the planar surface cannot be determined. No
other parameters of those in the namelist are needed to define the planar
surface.

If surface_shape = 'Cylinder' and coordinate_system =

'Cartesian' or coordinate_system = 'Cylindrical', then
the following parameters must be included in the namelist:

183
 top_cylinder_center_xyz,
bot_cylinder_center_xyz,
cylinder_radius

The function of the cylinder_radius real parameter is obvious. The

real 1D arrays top_cylinder_center_xyz and
bot_cylinder_center_xyz are of size 3. These are the (x,y,z)
coordinates of the centers of the top and bottom circular surfaces of the
cylinder, respectively. When coordinate_system =
'Cylindrical', only the z-coordinates in the two arrays are important
because the axis of the cylindrical surface is automatically aligned with the
z-axis. No other parameters of those in the namelist are needed to define
the cylindrical surface.

If surface_shape = 'Circle' and coordinate_system =

'Cylindrical' or coordinate_system = 'Cylindrical',
then flow is perpendicular to the circular surface and the following
parameters must be included in the namelist:

inner_circle_radius, outer_circle_radius,
surface_location, surface_coord_number

The function of first two real parameters is self-explanatort. The integer

parameter surface_coord_number identifies is the direction of flow
through the surface (i.e., it is perpendicular to the circular surface). Thus,
for flow along the x-direction, surface_coord_number = 1; for
flow along the z-direction, surface_coord_number = 3. The real
parameter surface_location identifies the location (coordinate) of
the circular surface on the axis specified by surface_coord_number.
Thus, if surface_coord_number = 3 and
surface_location=-40 m, then the circular surface is perpendicular
to the z-axis at this elevation. When coordinate_system =
'Cylindrical', only the surface_location is important (i.e., the
surface_coord_number parameter may be omitted from the namelist)
because the circular surface is taken to be perpendicular to the z-axis. No
other parameters of those in the namelist are needed to define the circular
surface.

If surface_shape = 'Sphere'and coordinate_system =

'Cylindrical' or coordinate_system = 'Cylindrical',
then the following parameters must be included in the namelist:

sphere_center_xyz, sphere_radius

The (x,y,z) coordinates of the center of the sphere are stored in the elements
of the real, 1D array top_cylinder_center_xyz of size 3.

184
 Of those three values, only the 3rd (corresponding to the z-coordinate of the
center of the sphere) one is used when coordinate_system =
'Cylindrical' because in this case the center of the spherical surface
is assumed to be at a radius r = 0. No other parameters of those in the
namelist are needed to define the spherical surface.

format_to_read_data:
This character parameter (up to 50 characters long) is needed as an input
when definition_mode ='NumberList' or 'NumberList', and
describes the format to read the list of number_of_connections
connection numbers or names. The number_of_connections
elements to be read according to the format format_to_read_data
are listed immediately after the end of the namelist (see examples in Figure
11.2).

NOTE: no other parameters of those in the namelist are needed if

definition_mode ='NumberList' or 'NumberList'.

The use of these inputs in the INTERFACES namelist is best illustrated in the
examples of Figure 11.2.

NOTE #1: It is possible to combine surfaces of very different geometries to create

interfaces that are very irregular in shape. Each one of these surfaces can be
defined independently, i.e., using different definition_mode values for each.

NOTE #2: All the various parameter combinations and permutations in the user-
supplied namelist Interface_Specifics must be defined for the simulation
to continue.

NOTE #3: All the surfaces within the interface must be defined. A total of
number_of_surfaces records (one for each surface of the interface) must be
provided in INTERFACE.2.1. Thus, if we define two interfaces, the first with 3
surfaces and the second with 4, then we need to provide the following data:
For the 1st Interface: One INTERFACE.2 record, and 3 SUBDOMAIN.2.1
records. For the 2nd Interface: One INTERFACE.2 record, and 4
INTERFACE.2.1 records.

The structure and use of the namelists in the data block INTERFACES is best
illustrated by the examples of Figure 11.2.

185
INTERFACES

&Interface_General_Info number_of_interfaces = 1 /
&Individual_Interface_Specifics interface_name = 'Int03',
number_of_surfaces = 1,
sign_of_flow_direction = 'DIR'
/
&Surface_Specifics definition_mode = 'NameList',
number_of_connections = 5,
format_to_read_data = '*',
/
'A00 4A00 5'
'A00 3A00 4'
'A00 5A00 6'
'A00 6A00 7'
'A00 7A00 8'

INTERFACES
&Interface_General_Info number_of_interfaces = 2 /
&Individual_Interface_Specifics interface_name = 'LBase',
number_of_surfaces = 1,
sign_of_flow_direction = 'DIR' /
&Surface_Specifics definition_mode = 'Geometry',
surface_shape = 'Rectangle'
Xmin = -2.0e3, Xmax = 5.0e03,
Ymin = -2.0e3, Ymax = 5.0e03,
Zmin = -1.1d2, Zmax = -1.1d2,
/
&Individual_Interface_Specifics interface_name = 'MBase',
number_of_surfaces = 1,
sign_of_flow_direction = 'DIR' /
&Surface_Specifics definition_mode = 'Geometry',
surface_shape = 'Rectangle'
Xmin = -2.0e3, Xmax = 5.0e03,
Ymin = -2.0e3, Ymax = 5.0e03,
surface_location = -3.0d1
surface_coord_number = 3,
/

INTERFACES
&Interface_General_Info number_of_interfaces = 1 /
&Individual_Interface_Specifics interface_name = 'Int03',
number_of_surfaces = 1,
sign_of_flow_direction = 'ABS'/
&Surface_Specifics definition_mode = 'Geometry',
surface_shape = 'Cylinder'
top_cylinder_center_xyz = 0.0d0, 0.0d0, -2.0d1,
bot_cylinder_center_xyz = 0.0d0, 0.0d0, -6.0d1,
cylinder_radius = 1.0d1,
/

INTERFACES
&Interface_General_Info number_of_interfaces = 1 /
&Individual_Interface_Specifics interface_name = 'Int04',
number_of_surfaces = 1,
sign_of_flow_direction = 'ABS'/
&Surface_Specifics definition_mode = 'Geometry',
surface_shape = 'Sphere'
sphere_center_xyz = 0.0d0, 0.0d0, -5.0d1,
sphare_radius = 2.0d1
/

Figure 11.2. Examples of the INTERFACES data block for tracking flows at specified interfaces.
This data block uses namelist-based formats for data inputs.

186
11.5. Data Block SS_GROUPS

This optional data block is a new TOUGH+ v1.5 feature that is unavailable to any earlier

version of the code. It allows monitoring the flow – instantaneous and cumulative –

through user-defined groups of sources and sinks (hereafter referred to as SSG). The time

series results are written in a number of separate output files that are equal to the number of

the SSG’s (one for each SSG), and are named according to the convention

SSGName_Time_Series, where 'SSGName' is the name of the SSG under

observation. The SSGName_Time_Series files are important in continuatin runs:

information from the last TOUGH+ simulation is gleaned from the older files and is used

to seamlessly continue the computations of the cumulative mass flows through the SSG’s.

If the keyword 'SS_GROUPS' is present in the input file, then the following data

are read using NAMELIST formats:

Record SS_GROUP.1

This record includes general data describing the number of interfaces to be

monitored. The namelist in this record is named SSGroup_General_Info and
has the general form:
&SSGroup_General_Info number_of_SSGroups = x /,

i.e., it contains only the following single parameter:

number_of_SSGroups:
An integer parameter describing the number of SSG’s at which time series
data are to be obtained. There is no upper limit in the number of SSG’s
that can be defined.

NOTE: The number_of_SSGroups parameter must be set to a positive

value for the simulation to continue.

The structure of the SSGroup_General_Info namelist and its use as the input
format in the data block SS_GROUPS is best illustrated in the example of Figure
11.3.

187
 Record SS_GROUP.2

The namelist in this record provides general information on each of the interfaces.
A total of number_of_SSGroups such records/namelists must be provided.
This namelist is named Individual_SSGroup_Specifics, has the general
form

&Individual_SSGroup_Specifics SSGroup_name ='x',

definition_mode = x,
number_of_SS ='x',
format_to_read_data ='x',/

and includes the following parameters:

SSGroup_name:
The interface name is described by this character variable of a maximum of
9 characters. Upon reading the variable, the TOUGH+ v1.5 code creates an
output file named SSGroupName_Time_Series, where
'SSGroupName' is the value of interface_name. If this variable is
omitted in the input, then the TOUGH+ code assigns internally the name
SSGroupName = 'SSGroupNN', where 'NN' is the 2-digit interface
number in the interface definition sequence.

definition_mode:
This character variable (up to 10 characters in length) describes the method
by which the region is defined. It can take the following possible values:
='NumberList': The region is defined by a list of connection numbers
='NameList': The region is defined by a list of connection names

number_of_SS:
An integer parameter specifying the number of individual sources and sinks
in the SS group. The number_of_SS parameter must be set to a positive
value for the simulation to continue.

format_to_read_data:
This character variable (up to 50 characters long) is needed as an input
when definition_mode ='NumberList' or 'NameList', and
describes the format to read the list of number_of_SS numbers or names
of the sources/sinks in the group. The number_of_SS elements to be
read according to the format format_to_read_data are listed
immediately after the end of the namelist (see example of Figure 11.3).

The structure and use of the namelists in the data block SS_GROUPS is best
illustrated by the example of Figure 11.3.

188
11.8. Warning Output and Error Messages
If inputs indicate conflicting conditions and/or parameter values are outside realistic

ranges, TOUGH+ is designed to respond according to the severity of the violation. Non-

critical conflicts result in a warning or clarifying message, internal adjustment of the

corresponding conditions and/or parameters, and continuation of the simulation. Serious

violations (e.g., initial conditions that violate fundamentals of physics and

thermodynamics) result in an error message identifying the problem and the interruption

of the simulation.

SS_GROUPS
&SSGroup_General_Info number_of_SSGroups = 1 /
&Individual_SSGroup_Specifics SSGroup_name = ' ',
definition_mode = 'NameList',
number_of_SS = 3,
format_to_read_data = '*'
/
'Pro01'
'Pro02'
'Wel06'

Figure 11.3. Example of the SS_GROUPS data block for tracking flows through specified groups of
sources/sinks. This data block uses namelist-based formats for data inputs.

189
 PAGE LEFT INTENTIONALLY BLANK

190
 Acknowledgements

This work was supported by the Assistant Secretary for Fossil Energy, Office of Natural

Gas and Petroleum Technology, through the National Energy Technology Laboratory,

under the U.S. Department of Energy, Contract No DE-AC02-05CH11231.

191
 PAGE LEFT INTENTIONALLY BLANK

192
 References

Americal Petroleum Institute (API), Division of Refining, Technical Data Book,

Petroleum Refining, American Petroleum Institute, Refining Department, Washington,
1977.

Barree R.D., and M.W. Conway, Multiphase non-Darcy flow in proppant packs, Paper
SPE 109561, 2007 Annual Technical Conference and Exhibition, Anaheim, CA, 11–
14 Nov 2007.

Bejan, A., Convection Heat Transfer, John Wiley & Sons, New York, 1984

Bird, R.B., W.E. Stewart, and E.N. Lightfoot, Transport Phenomena, New York: John
Wiley & Sons, Inc., 2007.

Brooks, R.H. and A.T. Corey, Hydraulic Properties of Porous Medium, Hydrology Paper
No. 3, Colorado State University, Fort Collins, Colorado, March 1964.

Cass, A., G.S. Campbell and T.L. Jones, Enhancement of Thermal Water Vapor Diffusion
in Soil. Soil Sci. Soc. Am. J., 48(1), 25 - 32, 1984.

Chung, T.H., M. Ajlan, L.L. Lee and K.E. Starling, Generalized multiparameter
correlation for nonpolar and polar fluid transport properties, Ind. Eng. Chem. Res.,
27(4), 671-679, 1988 (doi: 10.1021/ie00076a024)

Corey, A.T., The Interrelation Between Gas and Oil Relative Permeabilities, Producers
Monthly, 38-41, November 1954.

193
de Marsily, G., Quantitative Hydrogeology, Academic Press, Orlando, FL, 1986.

Doughty, C., Investigation of conceptual and numerical approaches for evaluating

moisture, gas, chemical, and heat transport in fractured unsaturated rock, J. Contam.t
Hydrol., 38(1-3), 69-106, 1999

Edwards, A.L., TRUMP: A Computer Program for Transient and Steady State
Temperature Distributions in Multidimensional Systems, National Technical
Information Service, National Bureau of Standards, Springfield, VA, 1972.

Falta, R.W., K. Pruess, I. Javandel and P.A. Witherspoon, Density-Driven Flow of Gas in
the Unsaturated Zone Due to the Evaporation of Volatile Organic Compounds,
Water. Resour. Res., 25(10), 2159 - 2169, 1989.

Fatt, I. and W.A. Klikoff, Effect of Fractional Wettability on Multiphase Flow Through
Porous Media, AIME Transactions, 216, 246, 1959.

Finsterle, S., iTOUGH2 User's Guide, Report LBNL-40040, Lawrence Berkeley National
Laboratory, Berkeley, California, 1999.

Finsterle, S., Implementation of the Forchheimer Equation in iTOUGH2, Project Report,

Lawrence Berkeley National Laboratory, Berkeley, Calif., 2001.

Forchheimer, P., Wasserbewewegung durch Boden, Zeit.Ver. Dtsch. Ing. 45, 1781, 1901.

Freeman, C.M., G.J. Moridis, and T.A. Blasingame, A Numerical Study of Microscale
Flow Behavior in Tight Gas and Shale Gas Reservoir Systems, Transp. in Porous
Med., 90(1): 253-268, 2011 (doi: 10.1007/ s11242-011-9761-6)

Fuller, E. N., K. Ensley, and J. C. Giddings, Diffusion of Halogenated Hydrocarbons in

Helium: The Effect of Structure on Collision Cross Sections, J. Phys. Chem., 73,
3679-3685, 1969.

Grant, M.A., Permeability Reduction Factors at Wairakei, paper 77-HT-52, presented at

AICHE-ASME Heat Transfer Conference, Salt Lake City, Utah, August 1977.

International Association for the Properties of Water and Steam (IAPWS), Release on the
IAPWS Industrial Formulation 1997 for the Thermodynamic Properties of Water
and Steam. Erlangen, Germany, 1997.

International Association for the Properties of Water and Steam (IAPWS). Revised
Release on the IAPWS Industrial Formulation 1997 for the Thermodynamic
Properties of Water and Steam, Lucerne, Switzerland, 2007.

International Association for the Properties of Water and Steam (IAPWS). Release on the
IAPWS Formulation 2008 for the Viscosity of Ordinary Water Substance. Berlin,
Germany, 2008.

194
International Association for the Properties of Water and Steam (IAPWS). Revised
Release on the Equation of State 2006 for H2O Ice Ih. Doorwerth, The Netherlands,
2009.

International Association for the Properties of Water and Steam (IAPWS). Release on the
IAPWS Formulation 2011 for the Thermal Conductivity of Ordinary Water
Substance. Plzeň, Czech Republic, 2011a.

International Association for the Properties of Water and Steam (IAPWS). Revised
Release on the Pressure along the Melting and Sublimation Curves of Ordinary
Water Substance. Plzeň, Czech Republic, 2011b.

International Association for the Properties of Water and Steam (IAPWS). Guideline on a
Low-Temperature Extension of the IAPWS-95 Formulation for Water Vapor.
Boulder, Colorado, USA, 2012.

International Formulation Committee, A Formulation of the Thermodynamic Properties of

Ordinary Water Substance, IFC Secretariat, Düsseldorf, Germany, 1967.

Itasca Consulting Group. FLAC3D: Fast Lagrangian Analysis of Continua in 3

Dimensions, Minneapolis, Minnesota, 2002.

Jones, S. C., A rapid accurate unsteady-state Klinkenberg parameter, SPE Journal 383–
397, 1972.

Jury, W.A., W.F. Spencer and W.J. Farmer, Behavior Assessment Model for Trace
Organics in Soil: I. Model Description, J. Environ. Qual., 12(4), 558 - 564, 1983.

Kim, J., and G.J. Moridis, Development of the T+M coupled flow-geomechanical
simulator to describe fracture propagation and coupled flow-thermal-geomechanical
processes in tight/shale gas systems, Computers & Geosciences, 60, 184-198, 2013
(doi: 10.1016/j.cageo.2013.04.023).

Klinkenberg, L.J., The Permeability of Porous Media to Liquids and Gases, in API
Drilling and Production Practice, 200–213, 1941.

Lam, S.T., A. Hunsbedt, P. Kruger and K. Pruess, Analysis of the Stanford Geothermal
Reservoir Model Experiments Using the LBL Reservoir Simulator, Geothermics,
17(4), 595- 605, LBL-25957, 1988.

Leverett, M.C., Capillary Behavior in Porous Solids, Trans. Soc. Pet. Eng. AIME, 142,
152-169, 1941.

Mason, E.A. and A.P. Malinauskas, Gas Transport in Porous Media: The Dusty Gas
Model, Elsevier, Amsterdam, The Netherlands, 1983.

Millington, R.J. and J.P. Quirk, Permeability of Porous Solids, Trans. Faraday Soc., 57,
1200-1207, 1961.

195
Milly, P.C.D., Moisture and Heat Transport in Hysteretic, Inhomogeneous Porous Media:
A Matric-Head Based Formulation and a Numerical Model, Water Resour. Res.,
18(3), 489 - 498, 1982.

Moridis, G.J., User’s Manual for the HYDRATE v1.5 option of TOUGH+ v1.5: A Code
for the Simulation of System Behavior in Hydrate-Bearing Geologic Media,
Lawrence Berkeley National Laboratory Report LBNL-6871E, August 2014.

Moridis, G.J. and C.M. Freeman, User’s Manual for the REALGASBRINE v1.0 option of
TOUGH+: A Code for the Simulation of System Behavior in Gas-Bearing Geologic
Media, Lawrence Berkeley National Laboratory Report LBNL-6870E, August 2014.

Moridis, G. and K. Pruess, TOUGH Simulations of Updegraff’s Set of Fluid and Heat
Flow Problems, Lawrence Berkeley Laboratory Report LBL-32611, Berkeley, CA,
November 1992.

Moridis, G. and K. Pruess, Flow and Transport Simulations Using T2CG1, a Package of
Conjugate Gradient Solvers for the TOUGH2 Family of Codes, Lawrence Berkeley
Laboratory Report LBL-36235, Berkeley, CA, 1995.

Moridis, G. and K. Pruess, T2SOLV: An Enhanced Package of Solvers for the TOUGH2
Family of Reservoir Simulation Codes, Geothermics, 27(4), 415 - 444, 1998.

Moridis, G.J., M.B. Kowalsky and K. Pruess, TOUGH+HYDRATE v1.0 User’s Manual:
A Code for the Simulation of System Behavior in Hydrate-Bearing Geologic Media,
Report LBNL-0149E, Lawrence Berkeley National Laboratory, Berkeley, CA
(2008).

Moridis, G.J., M.B. Kowalsky and K. Pruess, TOUGH+HYDRATE v1.1 User’s Manual:
A Code for the Simulation of System Behavior in Hydrate-Bearing Geologic Media,
Report LBNL-0149E, Lawrence Berkeley National Laboratory, Berkeley, CA
(2009).

Moridis, G.J., M.B. Kowalsky and K. Pruess, TOUGH+HYDRATE v1.2 User’s Manual:
A Code for the Simulation of System Behavior in Hydrate-Bearing Geologic Media,
Report LBNL-0149E, Lawrence Berkeley National Laboratory, Berkeley, CA
(2012).

Morrow, C., D. Lockner, D. Moore and J. Byerlee, Permeability of Granite in a

Temperature Gradient, Journal of Geophysical Research, 86(84), 3002-3008, April
1981.

Mualem, Y., A New Model for Predicting the Hydraulic Conductivity of Unsaturated
Porous Media, Water Resour. Res., 12(3), 513 - 522, 1976.

Narasimhan, T.N. and P.A. Witherspoon, An Integrated Finite Difference Method for
Analyzing Fluid Flow in Porous Media, Water Resour. Res., 12(1), 57 – 64, 1976.

196
Narasimhan, T.N., P.A. Witherspoon and A.L. Edwards, Numerical Model for Saturated-
Unsaturated Flow in Deformable Porous Media, Part 2: The Algorithm, Water
Resour. Res., 14(2), 255-261, 1978.

Pape, H., C. Clauser and J. Iffland, Permeability Prediction Based on Fractal Pore-Space
Geometry, Geophysics, 64(5), 1447 - 1460, 1999.

Parker J.C., Lenhard R.J. and T. Kuppusamy, A Parametric Model for Constitutive
Properties Governing Multiphase Flow in Porous Media, Water Resour. Res., 23(4),
618- 624, 1987.

Peaceman, D.W., Fundamentals of Numerical Reservoir Simulation, Elsevier, Amsterdam,

The Netherlands, 1977.

Peng, D.Y., and D.B. Robinson, A New Two-Constant Equation of State, Indust. and
Engr. Chemistry: Fundamentals 15, 59-64, 1976.

Phillips, O.M., Flow and Reactions in Permeable Rocks, Cambridge University Press,
Cambridge, New York, Melbourne, 1991.

Pickens, J.F., R.W. Gillham and D.R. Cameron, Finite Element Analysis of the Transport
of Water and Solutes in Tile-Drained Soils, J. of Hydrology, 40, 243-264, 1979.

Pruess, K., GMINC - A Mesh Generator for Flow Simulations in Fractured Reservoirs,
Lawrence Berkeley Laboratory Report LBL-15227, Berkeley, CA, March 1983.

Pruess, K., TOUGH2 - A General Purpose Numerical Simulator for Multiphase Fluid and
Heat Flow, Lawrence Berkeley Laboratory Report LBL-29400, Berkeley, CA, 1991.

Pruess, K., and T.N. Narasimhan, On Fluid Reserves and the Production of Superheated
Steam from Fractured, Vapor-Dominated Geothermal Reservoirs, J. Geophys. Res.,
87(B11), 9329 - 9339, 1982.

Pruess, K. and G.S. Bodvarsson, A Seven-Point Finite Difference Method for Improved
Grid Orientation Performance in Pattern Steam Floods, Proceedings, Seventh
Society of Petroleum Engineers Symposium on Reservoir Simulation, Paper SPE-
12252, 175 - 184, San Francisco, CA, 1983.

Pruess, K. and T.N. Narasimhan, A Practical Method for Modeling Fluid and Heat Flow in
Fractured Porous Media, Soc. Pet. Eng. J., 25(1), 14-26, February 1985.

Pruess, K., C. Oldenburg, and G. Moridis, TOUGH2 User’s Guide, Version 2.0, Report
LBNL-43134, Lawrence Berkeley National Laboratory, Berkeley, Calif., 1999.

Pruess, K., C. Oldenburg, and G. Moridis, TOUGH2 User’s Guide, Version 2.1, Report
LBNL-43134, Lawrence Berkeley National Laboratory, Berkeley, Calif., 2012.

Redlich, O. and J.N.S. Kwong, On The Thermodynamics of Solutions. Chem. Rev. 44(1),

197
 233–244, 1949. doi:10.1021/cr60137a013.

Riazi, M. R., and C. H. Whitson, Estimating Diffusion Coefficients of Dense Fluids, Ind.
Eng. Chem. Res., 32, 3081-3088, 1993.

Rutqvist J. and C.-F Tsang, A Study of Caprock Hydromechanical Changes Associated

with CO2 Injection into a Brine Aquifer. Environmental Geology, 42, 296-305, 2002.

Scheidegger, A. E., The Physics of Flow Through Porous Media, University of Toronto
Press, Toronto and Buffalo, Third Edition, 1974.

Sleijpen, G.L.G. and D. Fokkema, BiCGSTAB(m) for Linear Equations Involving

Unsymmetric Matrices with Complex Spectrum, Electronic Transactions on
Numerical Analysis, 1, 11 - 32, 1993.

Soave, G., Equilibrium Constants from a Modified Redlich–Kwong Equation of State,

Chem. Eng. Sci., 27, 1197-1203, 1972.

Somerton, W.H., et al., Thermal Behavior of Unconsolidated Oil Sands, Paper SPE-4506,
48th Annual Fall Meeting of the Society of Petroleum Engineers, Las Vegas, NV,
1973.

Somerton, W.H., et al., High Temperature Behavior of Rocks Associated With

Geothermal Type Reservoirs, Paper SPE-4897, 44th Annual California Regional
Meeting of the Society of Petroleum Engineers, San Francisco, CA, 1974.

Stone, H.L., Probability Model for Estimating Three-Phase Relative Permeability, Trans.
SPE of AIME, 249, 214-218, 1970.

TecPlot, Inc., Tecplot 10 User’s Manual, Bellevue, Washington, 2003.

Udell, K.S. and J.S. Fitch, Heat and Mass Transfer in Capillary Porous Media Considering
Evaporation, Condensation, and Non-Condensible Gas Effects, 23rd ASME/AIChE
National Heat Transfer Conference, Denver, CO, 1985.

van der Vorst, H.A., Bi-CGSTAB: A Fast and Smoothly Converging Variant of BiCG in
the Presence of Rounding Errors, SIAM J. Sci. Statist. Comput., 13, 631 - 644, 1992.

vanGenuchten, M.Th., A Closed-Form Equation for Predicting the Hydraulic Conductivity

of Unsaturated Soils, Soil Sci. Soc., 44, 892 - 898, 1980.

Vargaftik, N.B., Tables on the Thermophysical Properties of Liquids and Gases, 2nd Ed.,
John Wiley & Sons, New York, NY, 1975.

Vaughan, P.J., Analysis of Permeability Reduction During Flow of Heated, Aqueous Fluid
Through Westerly Granite, in C.F. Tsang (ed.), Coupled Processes Associated with
Nuclear Waste Repositories, 529 - 539, Academic Press, New York, 1987.

198
Verma, A.K., K. Pruess, C.F. Tsang and P.A. Witherspoon, A Study of Two-Phase
Concurrent Flow of Steam and Water in an Unconsolidated Porous Medium, Proc.
23rd National Heat Transfer Conference, Am. Society of Mechanical Engineers,
Denver, CO, 135–143, 1985.

Verma, A. and K. Pruess, Thermohydrologic Conditions and Silica Redistribution Near

High-Level Nuclear Wastes Emplaced in Saturated Geological Formations, J. of
Geophys. Res., 93(B2), 1159-1173, 1988.

Wagner W. et al., The IAPWS Industrial Formulation 1997 for the Thermodynamic
Properties of Water and Steam, ASME J. Eng. Gas Turbines and Power, 122, 150-
182, 2000.

Walker, W.R., J.D. Sabey, and D.R. Hampton, Studies of Heat Transfer and Water
Migration in Soils, Final Report, Department of Agricultural and Chemical
Engineering, Colorado State University, Fort Collins, CO, 80523, April 1981.

Warren, J.E. and P.J. Root, The Behavior of Naturally Fractured Reservoirs, Soc. Pet.
Eng. J., Transactions, AIME, 228, 245-255, September 1963.

Webb, S.W., Gas-Phase Diffusion in Porous Media - Evaluation of an Advective-

Dispersive Formulation and the Dusty Gas Model for Binary Mixtures, J. Por.
Media, 1(2), 187 - 199, 1998.

Webb, S.W. and C.K. Ho, Enhanced Vapor Diffusion in Porous Media - LDRD Final
Report, Sandia National Laboratories Report SAND98-2772, Albuquerque, NM,
1998b.

Wilke, C. R., A Viscosity Equation for Gas Mixtures, J. Chem. Phys., 18, 517-519,1950.

Wu, Y., Pruess, K., and P. Persoff, Gas Flow in Porous Media with Klinkenberg Effects.
Transp. Porous Media, 32, 117-137, 1988.

Xu, T., Y. Ontoy, P. Molling, N. Spycher, M. Parini and K. Pruess, Reactive Transport
Modeling of Injection Well Scaling and Acidizing at Tiwi Field, Philippines,
Geothermics, 33(4), 477 - 491, 2004.

Yaws, C., Chemical Properties Handbook, McGraw-Hill Education, 1999

199
 PAGE LEFT INTENTIONALLY BLANK

200
 APPENDIX

A Sample Input File

201
 PAGE LEFT INTENTIONALLY BLANK

202