CFDEM®coupling - User Forum

Hi all

I am attempting to simulate an unresolved coupled CFDDEM channel flow of particles using a fixed particle boundary for the walls instead of a meshed boundary. Currently I am running into an issue when using the READ_DATA command to bring in the wall particles, I have set the wall particles as type 2, and my flowing particles as type 1.

Hello all,
Is it possible to introduce two different liquid bridge forces in one simulation? As I see, for "easo/capillary/viscous" model, one can only use one liquid property for a simulation. How to use two liquid properties for two different class of particles?

Thanks in advance.

Hello,
Are there any known working Liggghts + OpenFoam containers out there on Docker Hub or elsewhere? I would like to use a container for testing out the solution as opposed to investing time in source compiles.

Thanks!
William

Hello all,

I am working on CFD-DEM coupling simulations with our in-house coupling solver, and we recently face a tricky technical problem. The problem is that our coupling simulation stops randomly without any errors when run the simulations in parallel.

In order to fix the problem, I have tried the following ways:

1. try to run the same simulations on three different machines (2 local workstations and one HPC cluster), however, all of them stop without any errors.

2. try to use different methods to decompose the computational domain, but still failed.

Hi,

I am new to this code and so far have been enjoying it.
I managed to get the two spheres tutorial case running for the immersed boundary solver.

My question is: Can I connect multiple spheres together to create a shape, give said shape material properties through the links between each individual sphere, and run the immersed boundary solver in CFDEM-public release?

Thanks
Andrew

Hi All,

I am new to simulations but a fast learner.

I installed ParScale (https://github.com/CFDEMproject/ParScale-PUBLIC) and am trying to run and visualize the example here (https://github.com/CFDEMproject/ParScale-PUBLIC/tree/master/examples/tes...).

Hi, I have problems compiling CFDEMCoupling on a Linux machine (Fedora34). I first had problems with VTK and compiling Liggghts. I used the auto makefile and switched auto install of VTK to on. However, I still had to set CFDEM_ADD_LIB_PATHS and CFDEM_ADD_LIBS in additionalLibs_5.x to get the compilation to work. I set the paths according to the following post: https://www.cfdem.com/comment/19083#comment-19083

setting：ErgunTestMPI
momCoupleModels: implicitCouple
forceModels: DiFeliceDrag, gradPForce, viscForce

In the script of LIG of ErgunTestMPI case: fix cfd2 all couple/cfd/force/implicit → it produces f_cfd2 and f_dragforce
probeModel: particleProbe → it produces DiFeliceDrag, gradPForce, viscForce.
IOmodel: trackIO → it produce DEMForces, impForces

What are the differences between f_cfd2, sum(f_dragforce), DEMForces?

Does the seed numbers used in initialization (in.liggghts_init) and run (in.liggghts_run) script require to be different? I am using initialisation script to generate the bed and run script to insert some particles with fluid on prepared gravel bed.

Can you please explain what is the difference between U, Us, and vSmoothField?