What
will be compared ?
Introduction
:
The purpose of this work is to compare, on a simple case,
Eclipse and Athos in order to put into relieves their respective avantages.
ATHOS and Eclipse are reservoir simulators able to satisfy all the basic reservoir engineering needs. It could be used to forecast a reservoir production as it will be in the example. Very powerfull, these two codes are very complex and allows to make very complete simulation .
Because it is impossible to compare the two codes in all
theirs particularities, the objective is to test the codes on a very simple
case and to help to choose the best code for the future generation of MFN
.
In order to be the most objective,
we will use only the pre and post processing devoted to the code .
Preprocessing
For Athos , a program which name is SIMView helps
the code to define the grid , initial cells and parameters of the simulation
. Because, Eclipse doesn't have an equivalent, the comparison will
be make without SIMView ( for the preprocessing ! ) .
The preprocessing is made by command lines for the two
codes . Thus, that will be these command lines which will be compared.
The facility to run the code and options they allow will determine the conclusion.
Here, the aim is to compare datas given by the codes and the software of visualisation of the parameters :
We want to simulate a reservoir
in a porous field. At the beginning, the field is full of oil. We inject
water in a cell and we recover the production of oil in another cell.
It can be interesting to display
:
About water :
- cell of injection (1,1,1)
- relative permeability
- viscosity
- density
- cell of production
- relative permeability
- viscosity
- density
- initially : full of oil
- 25 cells versus X. one versusY and Z
- meter
- cell of injection
- cell of production
- porosity : 0.2
- pressure at the top cell : 1000
- Permitivity versus X :
5 cells 10 5 cells 1 5 cells 10 5 cells 1 5 cells 10
- rock compresiblity : 1.E-10 ( neglected here )
The command lines are written in one file.
We can divide the data in different steps :
- GRID : internal parameters
- REGIONS : possibility to define different regions ( for a complex simulation )
- SOLUTION : other initialisations
specification of datas which will be written in files for the Eclipse Graphics Package
- SUMMARY :
precisions about :
- Schedule :
- cell injection
- cell production
- time steps of simulation
For the installation , it is only necessary to print in the .cshrc file :
setenv LM_LICENSE_FILE=/logiciels-mfn/Eclipse/macros/license.dat
To run : @eclipse file.DATA
( this file has been created by the preprocessing ).
The code creates different files :
The Eclipse graphics package is GRAF ; to run : @graf
First of all, you have to choice
a version of graf ( interactive with graphics one ) and to specify your
computer ( for Sun (Solaris 2) : 53 ).
|
|
|
Different steps with GRAF :
|
|
|
|
|
By this method, you can only plot graphs versus time .
If you want to visualize for example, the presence of oil in the grid and
to see the evolution versus time , let's read after .
It is easy to understand : the first lines open summary
, grid and resultsfile , then you create a new picture ( here pictuce 1
) , finally you create different graphs and you specify : grid, file
, XY ( or XZ or YZ ), 2D , results ( here soil: saturation of oil ) and
finally the step .
| menu 8-2 : execution of the intructions ..... | menu 5-1: display of the picture |
|
|
|
Picture obtained :
Command lines are written in a principal file ( .edo ) , other files ( .DAT ) where can be storage for example values of pressure versus depht and finally a file .sdo necessary for running .
These files allow to give physic parameters to the code without writting them in the .edo.
No run has been carried out .
The post processing is made by SIMView, software that can make Preprocessing.
First, in the .cshrc file :
Then, if you want to visualize an example of the demo , cope in your directory like this :
To run :
Because I didn't have enough time to install my results in SIMView ( it seems to be quite difficult ), I use the installation of former students ( Talbi Samia and Guenadou David for production of results and Samia Talbi for intallation in SIMView )
It become very easy to visualize exactly what you want
.... SIMView is well constructed and simple to use .
Examples of graphs :
First of all, I would like to insist
on the fact that this comparison is basis on a very simple test; it doesn't
take into account the powerfull of the codes; the aim is only to compare
the facility of codes in a first step.
|
|
|
|
| Preprocessing | written manual : + | separated files for initialisation : +
no written manual : + |
| Run |
|
|
| Postprocessing | direct integration of GRAF : +
user-friendly : - command line for 2D or 3D graphs : - Options for graph's presentation : - |
necessity to organize files for SIMView : -
user-friendly : + Options for graph's presentation : + |
On the one hand, by reading this table, Athos seems better
than Eclipse . For the preprocessing, the codes are quite equivalent but,
with a written manual, Athos is easier. For the run, I can not conclude
. For postprocessing, however SIMView needs a special installation of results,
it is more user-friendly, offers a lot of options, allows to plot 2D or
3D graphs without any difficulty ....
On the other hand, Eclipse is used by Schlumberger, ELF
and other big firms .... So , for a student, it can be
more interested to work on Eclipse than on Athos ...
However, because preprocessings are quite equivalent
and all results given by Eclipse are produced too by Athos, Athos is the
best choice .