[Volkanatar]

Hello Foamers,

I'm kind of new to OpenFOAM and conducting lagrangian particle tracking (lpt) simulations with the icoUncoupledKinematicParcelFoam solver.

I have a fully-developed velocity and pressure field which i get from a simulation with a steady-state solver in advance and take these fields of the last time step for the lpt simulation.

My problem is that when i run this lpt simulation parallel after decomposing the domain, the particles which are in the subdomain that belongs for example to processor0 are processed by the processor0 and so on as expected. So the particles are distributed to the processors based on their positions.

However i need that all the particles are distributed more or less evenly to the all processors. I think that kind of parallelization would speed up my lpt simulations.


Is that kind of parallelization possible for lagrangian field? I have tried to modify lagrangianFieldDecomposer.C (/src/parallel/decompose/decompose) file but it did not bring anything. In that case lagrangianFieldDecomposer.C this would be the right file to modify? Or is there any decompose method which is already doing that?

I would appreciate any help.

Thanks in advance

[Volkanatar]

Would anyone give me a hint about that?

So my problem is with load balancing: loss of performance for cases where particles are not uniformly distributed in the domain.

Thanks in advance.