Access to DEM data

MFiX
#1

Dear MFiX help,

I am planning to implement a particle agglomeration model on CFD-DEM results obtained from a MFiX-DEM simulation. Before implementing the model in MFiX, I would like to test its performance in another environment, e.g., MATLAB. In order to accomplish this, for the simplest model that we developed, I need to access to the following information:

• Particle ID, size & density
• DEM time step
• Center positions of particles at each DEM time step
• Velocities of particles at each DEM time step

Many of these information can be saved in the output .vtp files. However, I am not sure whether or not the writing interval can be adjusted as low as the DEM time step. When working with GUI, I feel that CFD time step has an impact on the writing interval for the saved DEM data (Output - Background IC - Particle data). Am I right?

According to the MFiX-DEM documentation, the DEM time step is calculated based on the values that we enter for the particle properties (size, density, normal spring constant, and dry restitution coefficient), DEM Δt = min(collision time/50). It will be nice to make sure that the DEM time step considered for the MFiX-DEM calculations matches with the value obtained from expressions provided in the MFiX-DEM documentation.

I do appreciate if you can advise me about the above-mentioned items and how I can access to the required information.

Thanks for your help, in advance.

Regards,
Jaber Shabanian

#2

The vtk files are typically written at a much larger interval than the fluid and dem time step. Otherwise the amount of data to postprocess becomes too large to handle.

The DEM DT is calculated in model/check_data/check_solids_dem.f. You can add a print statement to see its value. You can also look at the standard output and divide the fluid time steps by the number of DEM iterations to get the DEM DT (below DEM DT = 0.61728E-3 / 47=1.31E-5)

 Time =  0.69506E-02  Dt =  0.61728E-03
 Nit   HYDRO     Max res
 1   2.4E-03    V0
 2   0.3        P0
 3   0.2        P0
.
.
.
12   2.3E-03    P0
13   1.4E-03    P0
14   8.6E-04    P0
Timestep walltime, fluid solver:    0.062 s
DEM NITs: 47   Total PIP: 2400
 NITs/SEC = 3008.
Timestep walltime, DEM solver:    0.016

If you really want to write at the DEM time step scale, you will need to write some code in usr2_des.f and save the data you need in a format Matlab can read.

1 Like
#3

Thank you Jeff for the provided information.