Reservoir Simulators Comparison

[reservoirengineer]

Hello RE Experts,

Last time, Yeengineer started very interesting thread: Well Testing Software Comparisson.

[link Point to another website Only the registered members can access]
I am very keen to hear about your opinions for Reservoir Simulation software.

(1) Schlumberger Petrel/ eclipse family
(2) Landmark VIP Nexus
(3) CMG IMEX, Stars
(4) JOA/ Sensor
(5) Merlin
(6) Exodus
(7) etc.

I will not include Saudi aramco, Shell, Conoco, Chevron in-house softwares in this list because it wasn't available to purchase.

If you have to buy for your own use (company), what would be your choice? You will have to pay for the full amount.

I would like to hear about your selection based on
- technical aspects
- user friendliness
- cost/ benefit
- Simulation time
- industry standard
- reliability, etc.

Thanks.

[Shakespear]

I have been brought up on just one, ECLIPSE so my opinion is suspect to start.

Schlumberger has bought up the best peripheral software to build the simulation model run by ECLIPSE for the widest set of problems commonly run in the industry. PETREL being one of the the "recent" BIG purchases.

On the ease of use, well here I heard that SHELL supposedly had something good in this department. However this is hear say. ECLIPSE is a sophisticated tool and requires a high knowledge from its user. The peripheral programs also have fairly steep learning curves. Hence it takes at least a year to close the build cycle regarding using the software (I saw this with pure novices). To get to expert level will take many years and will depend on the type of problems you work on.

Development of ECLIPSE is done by top people that SLB can hire. I suspect they do not cut costs here. This is bread and butter for GEOQUEST.

ECLIPSE is frequently a requirement for Development Projects farmed out by oil companies to consulting firms. So it is a Industry Standard Simulator.

[jan]

I partly agree with Shakespear. I have worked above 10 years using Eclipse and 2 years using Petrel. According to me Eclipse has first place among simulators which were listed above. I utilized VIP, CMG, Merlin and Eclipse so far. Based on my experiences Eclispe is the best from them. Obviously work on Eclipse is a bit complicated but for people who know reservoir enginering and a bit mathematic it should not be a problem. I recommend Eclipse simulator with connection to Petrel.

[vergatario]

Gentelmen,
all the black oil simulators do the same basic tasks. The only difference is the way they handle the input and output information. May be Eclipse has several options that other simulators do not have, but it is because it has been an industry standard for long time and the options have been included by request and help of clients.
Regarding thermal simulator, it seem that Stars from CMG have better reputation than the others. E300, GEM and VIP are the compositional simulators that people use for compositional simulations. They work in the same way. First you find the EOS and later, run the simulation

[reservoirengineer]

Gents,

Thank you all for sharing your experiences. It is really valuable. I agreed some points with Shakespear for industry standard, functionality, etc.

But if you have a limited fund, will you choose Eclipse option to drain all your resources? It is a matter of resource allocation.
As everyone said, someone will need for Eclipse:
- years of experience to work on Eclipse model
- more than 1 dedicated reservoir simulation engineer to build model, QC, history match, etc.
- a single full field model how much time you will need to spend?

On the other hand, more user friendly, cheaper (non- Eclipse) simulator will accomplish some of the task. A single RE can do multiple full field simulations. And it will cost less than half of Eclipse?

I am sure accuracy is important but simulation is garbage-in/ garbage-out.

If it is exploration work, there is less data and you will have to work with analogous data and uncertainties.

Then what will be your decision?

[Shakespear]

Yes, I see what you mean. If you have limited resources then a cheaper simulator could be used for a "good enough" simulation. There is a lot of them around and one even comes with some books.

I once worked for a small company on a project where there were only 4 wells, and 2 were producers. Limited data and limited budget. Did I need eclipse? NO !!! I used a simple simulator for $400 and an a Nodel Analysis package to get results which were close enough the future production of the wells that were to be drilled. When the wells went on production the results were within +- 10%.

Maybe luck? Maybe simple reservoir? Don't know, but in this case Top of the line software would not have given any better answers in my opinion.

[reservoirengineer]

-----

[Shakespear]

It was something a friend of mine developed and then some company in France was trying to market :-)

Sh*t, I have forgotten what it was called as I used it a bit and then never came back to it again. Doubt if it is still around based on my GOOGLE search.

[reservoirengineer]

.----

[vergatario]

When you have limited resources then an option is to rent the simulator for the time scheduled for the simulation. There are some manager´s felling that you need to use a well known package to perform this important task. If you use a simulator that it is only known by few people then they will no trust in results.
The simulator that come with the Fanchi´s book is free. It is a modification of the Boast Simulator (DOE free software), which by the way has a package to
imput the data interactive and a postprocesor to see the output (you have to pay for it).
One good option it is to try to talk with new software provider (for with the Merlin, the new Coats´s simulator, exodus, etc,) in order to expand their market, that may offer attractive prices.

[reservoirengineer]

----

[temr]

Folks check ou this

[link Point to another website Only the registered members can access] I use this baby for almost a year almost ( less than 1% error) the same results with eclipse 100
Petrel like interface
Free parralel option
User frendly
Same data file keyword structure as eclipse uses