Reservoir Simulation

[reservoir_engineer]

hiiiiii all
i agree with all written from all brothers,
when i said yes you can work with eclipse without any help from petrel ,i was meaning that the static model can be imported from any other software like RMS, gocad,....i wasnt so clear.
for ACTNUM i think the Porosity may be good property for that but you have to care for the facies modeling for example shale values will be high and care also for the IOIP as you will remove volumes from the model.
i never use ACTNUM to reduce the no. of cells to decrease simulation time.
i always use UPSCALING module in PETREL to do this job.
you got two benifits
frist decrease the no. of cells without removing any part from the model,
second the upscalled grid will have average properties for the original ones,
at last you use well correlation window to check the validity of the properties of upscaled grids with the original properties for all wells.
i hope it is clear
Regards,
reservoir_engineer

[lesnew]

Dear All,
I am responsible for implementing a toll for optimization of production and exploitation of the Crude Oil and Natural Gas Extraction Plant using the data obtained from continuous measurement of phase contents in formation fluid. I shall appropriate production control and field management algorithms design and implement.
What you expect would be a level and span of the optimization?
Reservoirs, wells, pipelines, processing facilities, …?
Could you suggest something please?

Regards,
Chris.

[bratek]

Dear all,

Thank you very much!
Regards,
Bratek

[Shakespear]

i always use UPSCALING module in PETREL to do this job.

Yes, this is the way to go about it. If you have PETREL you should be doing all these adjustments here and not with ACTNUM. You also have to remember that ECLIPSE has certain keywords which were needed BEFORE powerful pre-processors were available to make the STATIC Model.

The Static Model should be built BY A GEOLOGIST unless you are in a small company with too much work and too few staff. The reservoir engineer should be involved as the geologist do make mistakes!!!!!!! That is why these are often team projects!!!!!

As a REng. you need to look at the logs and at least check how well cross-sections through your Static Model reflect what is seen at the wells. Are the shales where they should be, are the sands etc where they should be, is the volume properties around the well reflected in measure values from Well Tests, etc.

As for Seismic, again, a mistake in processing may put the horizons too deep or too shallow. Check, check and check again. Mistakes here could move the OWC in the wrong direction and give you a nasty surprise when they drill based on the model simulation results.

Bratek
I'm a REng so I can not directly address your question, however I would say this. As good as the software is today you will not do miracles. You will in the end chose what YOU THINK is the best method to work the data. The better you know the area the better job you will do.

If you are new in the area start asking questions of the geologist familiar with the area as every place is different. It also helps (in the case of PETREL) to know UNDER what circumstances SHOULD you use their different models to generate the properties based on your log/seismic data (garbage in is garbage out). Be conservative as you are dealing with something you DO NOT SEE and you have limited data.

Try to agree on some direction based through group discussion, group reviews. DO NOT BE intimidated not to say you do not know something. This very often this is a problem in our business.

[bratek]

Try to agree on some direction based through group discussion, group reviews. DO NOT BE intimidated not to say you do not know something. This very often this is a problem in our business.

WOW.......!
Many thank to Shakespear! ....... you are a WRITER engineering! your reply is so long which I have seen in this forum. .
Next steep, when I make upcaling model will need some more helps from you!

Because of lack of well data so the model of Horizon Variogam is not good ! (I don't know how to choose correct direction, range of the Horizon Variogram).

Now, I'm thinking about to make some Amplitude maps base on 2D seismic (!) which is maybe help me something ? ( I have never done it before)

Anyone have some experiences please help me more!

Thank you very much for your helps.
Regards,
Bratek

[Shakespear]

Thanks for the complement Bratek

Take a look here

[link Point to another website Only the registered members can access]
The discussion may give you some clues what to do. The following also had good points about QC of variograms.

[link Point to another website Only the registered members can access]

Variogram Panel

Variograms model the spatial continuity of whatever data is being gridded (in this case, subsea depth of formation top).

Click the Points button on the left of the panel to calculate the variogram points. The All Variograms graph shows the variograms for the northwest, north, northeast, and east directions as well as the vertical variogram. The well pair variance graph shows how the variance between wells varies with the distance between them. For those well pairs that are close (small separation distance) but that still have large variances, there is usually error associated with either the data, the location, or the markers. The areal variogram graph shows how reservoir continuity varies in the different directions.

Click the Curve button to fit a variogram model to the calculated points. If you don’t know which model to choose, select Spherical first. For structure depth type of data, the spherical variogram is ususally tried first.
Points to check in Variogram panel:

1. Make sure you have multiple wells if you want to calculate areal variogram.

2. You need multiple data points in each trace to calculate vertical variogram.

3. In the current version each project should contain one type of data for variogram calculation.

4. Calculate the variogram points first.

5. Curve fit the results. The fit is automatic as soon as you choose a model.

6. Other graphs may be brought up from the Graph button on the left. Change the graphs with Options pulldown menu Graphpar panel. For example you may turn off some of the curves on the All Variograms graph.

7. Variograms usually increases with distance.

8. Vertical variogram usually reaches high variance value in much shorter distance.

9. Well Pair Variance graph is a good indicator of data quality. For those well pairs that are in short distance but showing relatively large variance, there is usually problem in the data or well location.

10. Correlation Ratio to Vertical (CR_v) is usually larger than 10 (50, 100, 1000 depending on the data) if the vertical scale is depth. If vertical scale is layer index (when imported as layer data), it depends on the layer thickness and whether the data in different layers are correlated.

Go to the next panel by either clicking on the Next button on the lower left or by clicking on GridOption on the right hand side of the panel.

Although it is regarding Tops data, it is still useful to try and understand the logic.

Bratek, are you a geologist or a petroleum engineer? Or both Are you using PETREL?

PS: One of the papers mentioned in the above link was not there but I found it here

[link Point to another website Only the registered members can access]
At the above address you will find more papers

[link Point to another website Only the registered members can access]
This paper has your situation

Also available are interpreted well logs characterizing the vertical variation of porosity at eight different wells.

[link Point to another website Only the registered members can access]
Another good paper

[link Point to another website Only the registered members can access]
Well I am on a roll now, this paper you will find very useful

[link Point to another website Only the registered members can access]

[GerardoL]

Thank you very much !!!!
I m geologist. Two years ago I worked in a reservoir simulation with 1200000 cells, with 0.5 meters of thickness, and we worked with ACTNUM, the cutoff was 5 % of porosity. The field have large water injection and the history matching was very good !!!!!
The areal upscaling was in no productive areas of the grid.

best wishes

[Shakespear]

Just to clarify what I was thinking

You can always go into the DATA file and work with ACTNUM.

What I meant was that PETREL (I believe as I am not a user of it and it has replaced FLOGRID which I use) should have the option for you to set cut-offs or conditions to make cells inactive and then translate this into ACTNUM or other keywords when it outputs a file to be used by ECLIPSE ( model description in terms of properties and grid geometry).

Same would need to be done for cells which have tiny pore volumes which could create problems running the model.

[indianoilman]

hey Shakespear.... how to use flogrid in place of petrel! Please...

[Shakespear]

OK, here we go.

PETREL was bought by Schlumberger as it is a top of the line application for modeling the geology. Hence, it is an application which FIRST in line is made for a geologist, and only after this I would say for a reservoir engineer. Unless you have a simple situation or work in a small company, the Static Model will be made for you by a geologist!!!!

FLOGRID is made for a reservoir engineer. He would go to the geologist and ask to be given logs, TOP/BOT maps of horizons etc., and make the Static Model himself. I am simplifiing what I say above but in a nut shell that is what happens. FLOGRID is not primarily for geologists.

How to use FLOGRID? Well the simplest way to learn it if you have no one around to show you is to get the Data Files for the tutorials and go step by step to try to understand the logic of the problem. Understanding the logic is the first hurdle. The second is learning where various options are and how/when to use them!!!! THIS TAKES TIME, thus work on something simple until you understand it. (The same goes for SCHEDULE program!!!!) It is easy to screw up!!!!!!

Last is data preparation (or maybe it should be first). The format must be right, it should be input into FLOGRID correctly and then DOUBLE CHECKED!!!!!!!!!!!!!!!

PETREL and FLOGRID are able to record the loading/processing steps when building a model. Thus if some data changes you just run the command file and FLOGRID will automatically rebuild the STATIC Model

So if this is the way you want to go get someone to upload the DATA files for Tutorials and start there.

[GerardoL]

Hi friends,
It is very very very good book of Luca Cosentino.
He talk about everything, data, modeling, simulation, etc.

[link Point to another website Only the registered members can access]

[indianoilman]

THANK YOU ALL FOR SUCH AMAZING DISCUSSIONS...!
i'm a reservoir engineer trying to learn reservoir simulation and use of eclipse...