Re: [Sgems-users] reservoir simlulation model

[Jose R. V. <vi...@gm...>]

Eugenio,

One simple and popular way of doing that is to simulate porosity using
Sgsim in the full high-resolution fine-grid model. Then, you simulate
permeability using collocated porosity with CoSgsim (use Markov model
1 for cokriging option and provide the correlation coeff between
porosity and permeability). Repeat this process for each facies
category (e.g. sand and shale) and finally condition the simulated
porosity and permeability values by facies using the "cookie-cut"
procedure, i.e. visit each facies gridblock and find out the
(snesim-simulated) facies category and them pick the corresponding
porosity and permeability. You can set up a Python script in SGeMS to
do this. I have a Matlab script I wrote some years ago but you might
want to perform this process in SGeMS. Most likely, the SGeMS
community can provide such a python script for you (I would also be
happy to acess it).

That simple cookie-cut procedure is actually inefficient but it is
used (maybe widely) in industry. Keep in mind that discontinuity will
be "seen" at the facies boundaries that might not be expected in
reality, but obviously this is an approximation.

Snesim is used to generate facies models based on TIs. Alternatively,
Filtersim could be used for categorical variables but also for
continuous variables. You might want to have a look at the Filtersim
algorithm.

Hope that helps you as a first approximation to solve your problem.

Kind regards,

Jose R. Villa

On 3/4/09, Eugênio Silva <eug...@ho...> wrote:
>
>
>
> Dear list,
>
>
>
> when the SGeMS generates a realization by SNESIM algorithm, the result of
> the simulation is a qualitative description of the grid. In this
> description, each grid block is identified by a value and this value
> represents some facies of the reservoir. For instance, if there are two
> facies in the grid, each grid block is identified by 0 (absence of a
> permeability channel) or 1 (absence of a permeability channel). However, if
> I want to create a IMEX or ECLIPSE simulation model, I need to generate a
> PERM section in a DAT file with the permeability values in each grid block.
>
>
>
> My doubt is: how can I transform the SNESIM qualitative description in a
> quantitative information of permeability?
>
>
>
> Regards,
>
>
>
> Eugênio.
>
>
>
>
>
>
>
> _________________________________________________________________
> Conheça o Windows Live Spaces, a rede de relacionamentos do Messenger!
> http://www.amigosdomessenger.com.br/


-- 
Jose R. Villa