Which MRST model is best for reduced-order modeling (POD/MOR)?

[TEBIB Walid]

Dear MRST developers and users,

I am new to MRST and would like to experiment with model order reduction (MOR) techniques such as POD to speed up reservoir simulations.

Could you please advise which MRST model or simulator (e.g., black-oil or AD-based models) would be the best starting point for this type of work? I am especially interested in models that are easy to extend and allow access to simulation states and equations.

Any pointers or example recommendations would be greatly appreciated.

Best regards,

Walid

[Olav Møyner]

Hi,

The ad-based solvers (ad-core/ad-blackoil) are probably the most relevant for this. There is a lot of details in the free MRST books on how they work: https://www.sintef.no/projectweb/mrst/documentation/

The deprecated ad-fi solvers that are the predecessors to the ad-core solvers contains a rudimentary POD implementation. The code for the POD itself could be worth a look, even if the solvers are quite old and slow (and will not be updated any more): https://github.com/SINTEF-AppliedCompSci/MRST/tree/main/autodiff/ad-fi/orthodecomp

https://github.com/SINTEF-AppliedCompSci/MRST/blob/main/autodiff/ad-fi/examples/experimental/pod/simplePOD.m

Best regards,

Olav

--
You received this message because you are subscribed to the Google Groups "MRST-users: The Matlab Reservoir Simulation Toolbox User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sintef-mrst...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sintef-mrst/ecaefd63-081f-4998-801e-60e7b03558bfn%40googlegroups.com.