Reservoir Simulation

[aliali]

Can any1 explain to me the main idea behind reservoir simulation ?
i mean what shall i study , which softwares will be at help, am trying to work in this field when i graduate so i wanna prepare myself ,so what do u suggest ?

have anice day

[aliali]

no1 knows about it or what ?

[pump2000]

resevoir simulators are just tools for reservoir engineerings
they are programs to solve numerically flow egution of fluid in porous media
the most Known one is eclipse 100 for black oil and 300 for compositional
you can start learning about reservoir simulation by reading the first published book about by KHaled aziz 1977 I am not sure about the date
it is old and very good starting point about the theory behind that
but if want to learn how to simulate I mean the practical side you have to have special training courses of shlumberger which they are very very expensive but however u can download the program and the ***** find simple cases(very well kown one to caliprate the simulator infact try SPE cases) to run and try to understand them and sure the manual of eclipse well help you alot the technical manual,
I will give an idea about Eclipse later
see

[reservoir_engineer]

al salam 3alikom
i read the reply from pump2000, it is very good informations,
reservoir engineering based on diffusivity equation which can be solved by many methods, analytical methods that used for well testing and aquifer description.
reservoir simulation is numerical solution for the diffusivity equation.
reservoir model consist of two phases static phase( structural description for your reservoir and it is the task of the geologist) and the dynamic phase is to load the cells with the fluids and initiate your model with the wells that are matching points for the model in RFT data, production data (BFPD, BOPD,WC, WHP, BHFP,....)
after you match these parametres for all wells you have to run many sensitivity cases to show the response of your model for the change in the uncertain parameters, then you can use your model for prediction.

[bratek]

al salam 3alikom
i read the reply from pump2000, it is very good informations,.

Can you upload some reservoir simulation documents?
thank you very much!

[hamdi1988]

reservoir simulation is creating fluid flow equation that describes the reservoir performance

[indianoilman]

is it possible to use eclipse without taking help from petrel... if yes ..then how to create static model... and start using eclipse..?

[reservoir_engineer]

hiiiii all
yes it is possible to work with Eclipse without any help from Petrel.
any modeling software (static or dynamic state) must be able to import to and export from other softwares.
for example Petrel can export static model to VIP, CMG, Eclipse, timpest,....
the same CMG, Geoframe, Landmark, RMS,...
it dependinding on the available static software you have.
i hope it is clear,
Regards,
reservoir_engineer

[bratek]

is it possible to use eclipse without taking help from petrel... if yes ..then how to create static model... and start using eclipse..?

There are two models for modeling:
1. Static modeling by using (Petrel, Gocad, RMS,...)
2. Dynamic modeling by using (Eclipse, CMG, VIP,...)

So we can not run dynamic modeling (simulation) without Static model !

----------------------
Fist, we need Petrel for built (1) 3D structure model (hoizons/faults) --> (2) Facies modeling --> (3) Properties modeling (K, Phi,...) --> (4) Upscaling model (for simulation) --> (5) export model.

Second, we import (static model from Petrel) to Eclipse and then run simulation,....... and so on ( I have never used Eclipse)... please some one give some more ideas here....

So the result of simulation modeling is also depend on STATIC model ( Geostatistic model) from Petrel
----------------------
Some one have other ideas or suggestions please discuss more.!

Thank you!
Regards,
Bratek

[GerardoL]

I agree with bratek, very clear !!!
When you run the simulation and the value is the same that real production (history matching) then you can to make the forecast for the future production

If the simulation no represent the real production you need modify the STATIC MODEL (with Petrel or others).

The model of the secondary production (with water injection ) is more difficult that only primary production. Other important simulation is with stream lines.

good luck

[Shakespear]

The steps are as per Bratek.

The STATIC model describes the geometry of the field and the flow layer properties ( perm, porosity, net thickness, depth, etc).

ECLIPSE needs this Static model plus PVT properties to model the fluids present, relative permeability to model what happens to perms for the different phases as saturation changes, capillary pressure curves to model distribution of fluids in your dynamic model, location of the contacts (oil-water, gas-oil, etc depending on what fluids you have initially in the reservoir), aquifer if present. These are the big ones for getting started.

Once you get the model together, the first question will be if it is giving you the right volumes of gas and oil as compared to you volumetric calculations. This validates that you have the HC in the modeled right !!! If not you will need to start to check things that effect volumes (PVT, Static Model, contacts, capillary pressure curve etc).

Assuming volumes match, now you run the model and see if the models the gas/oil/water production on a field, region or per well level. This is where the fun starts and experience becomes important.

[GerardoL]

If you use all cells, the time of the run simulation es very long.
In Eclipse we use property in orden to decrease the simulation time called ACTNUM (ACTNUM=0 inactive cell and ACTNUM=1 active cell). This property has a cutoff value.

ACTNUM is definited in the STATIC MODEL and exported a Eclipse.

What property do you use for ACTNUM (i.e. porosity, SP distributed with Kriging, etc.)?