[celestial]

Hi

I am referring to my recent post

Discontinuous conductivities in heat transfer

From the lack of responses, I conclude that there are no block coupling solvers involving fields over adjacent meshes.

And even after reading code from

blockLduMatrix

I really don't know where to start if I have to do it on my own.

Therefore, this thread is a request to add such solvers to open-foam-extend.
I am sure there are individuals in the OpenFOAM community that are much more capable than little me and who could handle this task.

It would be greatly appreciated since I have hit a roadblock.


Regards to all

[celestial]

I'll give you a hint though.

The forebear method ( a hack that I believe was specific conjugate heat transfer only)

is to be found on p 10 of

Implicit solution techniques for coupled multi-field problems – Block Solution, Coupled Matrices

Advanced Training at the OpenFOAM Workshop Jun 6 2010


and, the generalized version is to be found on p 43 of

Block-Coupled Simulations Using OpenFOAM

6th OpenFOAM Workshop 13-16 June 2011


Therefore, probably the latter approach is the way out, together with the appropriate boundary condition that is desired at the interface between the meshes. So I think that after more than 3 years, I finally see light at the end of the tunnel.

But to be forthright, I'm still surprised that nobody shows any interest in this kind of block coupling.

Anyhow, thanks for viewing for my posts.

[dlahaye]

The lack of response might be related to the complexity of the problem that you raise.

In your post you discuss both block-coupled solvers and multi-region problems. Both these issues have recently attracted a lot of attention in this conmunity recently.

For multi-region problems, e.g., implicit handling of coupled patches have been proposed in https://www.openfoam.com/news/main-n...mplicit-cyclic.

Block-coupled solvers have discussed in e.g. the recent OpenFoam workshops.

A good overview of both topics might be required to make a next step.

I am happy to share my limited insight.

[celestial]

Sorry but I fail to see why a simple mathematical problem should be so difficult to implement. But it does not matter since I am closing the thread as I was using the wrong terminology all along:

Implicit solution techniques for coupled multi-field problems – Block Solution, Coupled Matrices

Advanced Training at the OpenFOAM Workshop Jun 6 2010

discusses two topics concerning implicit coupling namely

Domain (Matrix)

and

Equation (Block)

coupling

What I am interested in is domain coupling ( so one variable in my case over
more than two adjacent meshes)

The hack for this was inserted into conjugate heat transfer solver long time ago - see p 10 in the above pdf where you will find a function

hook part of the coupledFvScalarMatrix class

This function was later renamed set as I checked by downloading the source code for foam-extend-1.6 and that code was later improved and expanded upon in later foam-extend versions.

So unless I am wrong, I have a feeling one can add the boundary conditions

k1*grad T1 = k2*grad T2 (adjacent meshes with different heat conductivities)

to the overall matrix equations

which would mean that I could have solved this problem more than 3 years ago with existing OF code instead of wasting my time


Such is life