modelling 3000 fractures in porous media matrix

[Oops-Qiao]

Hello,

I just have a quick question regarding the Hydraulic-thermal simulation of 3000 fractures in porous media matrix.
When I checked the moose website, I think it is suggested to use MultiApp to couple fractures (one subApp) and porous media (Main App). Am I right? Will it be very slow for computation?
I just read another paper which mentions that their approach can accelerate the coupling between fracture and matrix (https://arxiv.org/pdf/1709.02493). What do you think about it? Can I code their approach into moose?
Thank you in advance.

Best regards,
Qiao

@WilkAndy @GiudGiud

[WilkAndy]

Hi @Oops-Qiao . I suspect that your problems won't be in the speed of the transfer between Apps. In my experience, the latency from transferring data is unnoticeable. Yes, you could code that paper's idea into MOOSE. But, you'll probably have more problems with meshing and with unexpected things that occur when you embed a 2D mesh into a 3D mesh, as documented in the webpages you've obviously read (for future readers: https://mooseframework.inl.gov/modules/porous_flow/multiapp_fracture_flow_introduction.html ). So, i suggest just try with what's currently available in MOOSE before going to a lot of effort to make something new.