I want to make sure I understand the mathematical implementation of the no-slip, partial-slip and full-slip boundary conditions in a pure DEM framework (no fluid) with a linear spring-dashpot contact model.
I’ve noticed that the friction coefficient for particle-particle and particle-wall interactions is input in the “Solids” dialog (see below) as a property of the particles, while the BC dialog has options for “no-slip”, “partial-slip”, and “full-slip” at a wall boundary. The description for what these different boundaries mean are in terms of the Johnson-Jackson equations for TFM, which is (or should be) irrelevant for a pure DEM solver. I want to make sure I understand what these selections in the GUI do for pure DEM.
I assume “no-slip” means ignore the Solids dialog input for kinetic friction coefficient (i.e. just tangential spring-dashpot model), “full-slip” means ignore the tangential spring-dashpot model and just use kinetic friction, and partial-slip would be a hybrid of the two? I’ve included what this would look like mathematically below.
Please let me know if this isn’t correct! I couldn’t find a description of these BCs in the documentation so if I just missed it, I’d love to know that as well!
If you are working on pure DEM, my understanding is that you only need to set those parameters (fraction, restitution …with wall) in the Solids dialog. The solid wall boundary type in the BC dialog (no matter what you choose no-slip, free slip, partial slip) will not have any effect. You only need to let the code know that there is a wall. The boundary type (no-slip, free slip or partial slip wall) are required for fluid phase (or fluid-like solid phases), actually should be combined into “wall” when pure DEM is activated.
No, the CFD domain should be equal to or larger than the DEM domain (you can limit the particles in part of the CFD domain by UDF.) Not clear what condition you are going to simulate when you say the CFD domain is smaller than the DEM domain.
You can find papers that talk in more detail about why your DEM particle size must be smaller than your CFD cell size. One good place to start is “Challenges of DEM: I. Competing bottlenecks in parallelization of gas–solid flows” which goes into choosing appropriate simulation conditions in MFIX.
Thanks very much for your valuable comments. The ultimate aim I want to realize is applying pressure on particle assembly during the simulation. The following post of mine discuss this topic preliminary.
The TFM-DEM hybrid is not supported and is in no way guaranteed to work. There is code in the solver to run hybrid simulations, and there is also code present in the GUI, which is disabled. We do not encourage creation of new TFM-DEM hybrid models but the GUI will automatically enable the Hybrid mode if it finds a Hybrid model.
To access this “hidden” feature - create a new case from the “blank” template. In the “Model Setup” pane, Set “Solver” to “TFM”. In “Solids”, create 2 solids phases. Then save the file and exit.
Now, in an editor, open the .mfx file and change the solids_model for the second solid from TFM to DEM. Save the file and restart MFiX. You will now have the Hybrid mode enabled.
This is not guaranteed to work and we can only offer limited support for the hybrid model at this time - it is not under active development.
