DisplacementDeltas not implemented for explicit coupling?

[uekerman]

https://github.com/precice/calculix-adapter/blob/master/adapter/PreciceInterface.c

It seems like sim->coupling_init_v is only updated for implicit coupling (through Precice_WriteIterationCheckpoint), but never if only explicit coupling is used.

@arusch Are we overseeing something?

Thanks!

[nkr0]

Is there a difference in how Precice_IsWriteCheckpointRequired responds in implicit coupling and explicit coupling?

calculix-adapter/nonlingeo_precice.c

Line 1116 in 10938d8

if( Precice_IsWriteCheckpointRequired() )

[uekerman]

Yes, for explicit coupling, it will always return false since you don't need checkpoints. For implicit coupling it returns true in the first iteration of a timestep (that's when you need to store the checkpoint) and false afterwards.

[nkr0]

So, when should sim->coupling_init_v be updated in explicit coupling? Do you mean for simulation restarts?

calculix-adapter/adapter/CCXHelpers.c

Line 83 in 10938d8

coordinates[i * 3 + 0] = co[nodeIdx * 3 + 0] + v[nodeIdx * mt + 1];

Here the previous solution is added to coordinates and the comment says it is for restarts. Although I'm not really sure how correct that is. Displacement is the solution from a loaded condition. But changing the coordinates really alters the solid. It can stay in the displaced state without any load acting on it.

Possibly there are other parts to the restart setting that I haven't seen.

[uekerman]

No, this should not be related to restarts. For explicit coupling, sim->coupling_init_v needs to be updated after every timestep.

[nkr0]

Right!

calculix-adapter/adapter/CCXHelpers.c

Line 148 in 10938d8

displacementDeltas[3 * i] = v[nodeIdx * mt + 1] - v_init[nodeIdx * mt + 1];

This must be why I was getting wrong velocities when computed at the fluid side from DisplacementDeltas in explicit coupling. I thought the mistake was on the fluid side and added functionality to transfer velocities.