PFLOTRAN (www.pflotran.org) is actively maintained and supported, written in modern Fortran 2003, has an extensive set of capabilities, and runs on machines ranging from laptops to the largest-scale supercomputers.

I have heard of http://www.openfoam.com/, but I do not know if it works for porous media! \m/

@Eduardo: OpenFOAM is a C++ toolbox generally used for *build* solvers. It has much broader application and therefore is not helpful as a ready-to-use reservoir simulator. Thank you for the answer.

@Nishikant: My Google search for "Finchi IFLO simulator download" wasn't successful, the closest match: http://www.fanchi.com/IFLO.htm

Could you please share the link for the sources? It seems IFLO isn't a FOSS, the author only mentions the executable in that website.

Thanks.

I would neither say that Dumux is at an early nor that it is unmaintained. It is based on DUNE http://www.dune-project.org which started in 2002, is well maintained, and has just released its version 2.2.1. Take a look at the pdelab module www.dune-project.org/pdelab/, which we use a lot for porous media flow.

There is also another effort by academia and industry based on DUNE. The Open Porous Media Initiative http://www.opm-project.org, which is already used for reservoir simulation and CO2 storage (just like dune-pdelab). It will release version 1.0 this month.

Disclaimer: I am one of the DUNE maintainers and also involved in OPM

@Markus: I managed to build DuMuX some time ago and my first impression was that of a collection of C++ modules being used to describe particular scenarios remounting reservoir engineering, not a fully featured reservoir simulator. I have to admit i didn't explored DUNE carefully nor the examples in the bundle.

It's great have you here as one of the maintainers, i would like to ask you:

1. What exactly is the relation between DuMuX and the OPM initiative? Is the initiative considered software?

2. There are plans to integrate all those disconnected modules to produce an executable? My conception of a simulator is pretty much that of CMG (IMEX, STARS, ...) We describe the problem in a text file *.dat and the executable parses it producing the results *.out.

I'll reserve time to build DuMuX/OPM again and see how everything is organized.

Thanks.

@julio: Your original post was about flow in porous media and not mentioning reservoir simulation in particular. Unfortunately, I am not a reservoir engineer and therefore have not worked with any commercial reservoir engineer simulation software before and cannot compare our approach with CMG based on first hand experiences.

From my limited point of view IMEX is three phase black oil simulator, and Stars is a thermodynamical simulator. Both can be produced with dune-pdelab, DuMux and OPM. If your are tied to a specific output/input (apart from eclipse data file) format, you will need to program a little. In your case I would go for OPM as it will probably be used by more people from your field.

In general the assumption used by people in the financial industry holds for software, too: "There is no such thing as a free lunch". In our case this means: Either you use proprietary simulator, pay good money, and have to live with the features/formats that are there. Or use open source and can control exactly how your simulator works, and are able to implement the features you are missing yourself, hope for others to do it, or pay others to do it. Personally (as a scientist), I prefer the latter approach. I am biased of course but I know several commercial users that are starting to share this view (e.g. because commercial alternatives produce wrong results for some methods or a not fast enough).

Let me answer your questions:

1. OPM is based on DUNE and partly reuses some of the code of DuMuX (e.g. in ewoms). It's aim is to advance simulation tools for fluid flow and transport in porous media targeted for the oil industry. I would call it a software framework that gives you the ability to build your own custom simulators (see the examples provided with the source code).

2. There are examples of how to build such things in the various modules (opm-porsol: semi-implicit pressure and transport solvers using the IMPES method, opm-upscaling: single-phase and steady-state upscaling methods, ewoms: fully-implicit solvers for flow and transport in porous media, including blackoil). Based on this one can implement one's own custom simulators.

COMET costs very little and is quite good, if it is only a budget issue.

@Barry: Free as in freedom, having the source code makes a huge difference.

AH - well Sintef have a source code, but it uses Matlab as an engine...

There's also the code from CSIRO, SIMED II (Coal Seam Methane), OpenGeoSys are doing something...

@Barry: I'm bookmarking all for future assessment, thanks.

@barry Sintef is also a contributor to OPM. Parts of their matlab code is replicated there.

Depending on whether you want to simulate Darcy type of flow, which was answered by most above, or pore scale flow.

For the latter, there are some simulators, e.g. computing permeability from micro-CT images, like Palabos http://www.palabos.org/

OpenFoam http://www.openfoam.com/

OpenLB http://optilb.org/openlb/

Thank you @Steffen for adding to the list.

PFLOTRAN (www.pflotran.org) is actively maintained and supported, written in modern Fortran 2003, has an extensive set of capabilities, and runs on machines ranging from laptops to the largest-scale supercomputers.