Coupling models in LIGGGHTS?

Submitted by yuanjl on Mon, 02/18/2013 - 09:27

Hi everyone,

I am just coming back with a new equiry on CFDEM coupling.

We are doing modeling and simulation of transport processes in the energy systems, and the physical domain consists of: 1) the porous layers with micro or even nano meter scaled pores/solid particles, 2) a open channel (for the fluid flow) surrouded by the solid plate, whose size (in mm) are much bigger than the porous layers to supply the reactants to the porous layers. The transport processes include the heat transfer (convection in the channels and pores, and conduction in the solid particles), mass transfer in the channels/pores, which are affected by the reactions on the solid particles in the porous layers.

I am just thinking if CFDEM coupling is able to be applied for this case? For instance, we may generate a box for the porous structures by some way, such as by LAMMPS or similar, and use CFDEMcoupling for the transport processes in the porous layers, and pure CFD for those in the non-porous regions. Is it feasible?

If it is positive, which method is the best for the porous structure construction (to represent the discrete solid materials and pores), in terms of data loading/exchange and inplementation in LIGGGHTS? As supposed, we are allowed to use multi-functions (or mudules) of the LAMMPS in the LIGGGHTS at the same time? Another question is that if it is possible to put the solid particles as the fixed ones, while the fluid in the pores are set as movable particles representing multi-compoent gas mixture?

All the comments and suggestions, even the references, are most welcome.

Thanks very much in advance...

ckloss's picture

ckloss | Tue, 02/19/2013 - 00:01

Hi yuanjl,

many questions... If the question is if you can do online multi-scale simulations of the processes at the porous nano scale layer coupled to a macroscopic CFD simulation, the answer is yes, but not out-of-the-box. You'd have to develop a mapping from the coarse CFD simulation to get the boundary conditions of the nano scale simulation, and you'd have to implement the back-mapping of the macroscopic properties.

Sandia has done something similar in another context:
http://lammps.sandia.gov/doc/fix_atc.html

Cheers
Christoph