Hi,

As shown below, the volume fraction distribution (i.e. the left one )of PIC simulation doesn’t looks smooth as the mole fraction distribution of gas specie (i.e the right one) does. Is it possible for us to improve its resolution?

Thank you!

Hi,

As shown below, the volume fraction distribution (i.e. the left one )of PIC simulation doesn’t looks smooth as the mole fraction distribution of gas specie (i.e the right one) does. Is it possible for us to improve its resolution?

Thank you!

The parcels are being binned on the fluid mesh to determine the volume fraction. So what you see are the fluid cells. Not sure what your goal is:

- Higher fidelity simulation
- Better looking image

If it is Higher fidelity simulation, increase your cell count. If you want a better looking image, you can visualize the parcels instead, or use various interpolation schemes to smooth out the image.

Hi, Weber

What i confused is that why the gas component mole fraction distribution (the right one in the figure above) looks much smoother than the gas volume fraction distribution (the left one in the figure above), both of which are post-processed by ParaView. As you see, my bed geometry is irregular, is there an easy way to improve its resolution by interpolating. If this function is added into MFIX GUI, i think it will be very helpful for MFIX-PIC user to get high resolution image.

On the other hand, as you suggested, i tried to visualize the parcels in ParaView. Another question is how to obtain the real image of the particle distribution. You know, the sparseness of the particle distribution is related to the scaling factor of the parcel size.

Is the visualization in the GUI not good enough?

Hi, Weber

Thank you for your timely reply. My case ran on the supercomputer, and there is no GUI. Therefore, a convenient way to postprocess the result is by using ParaView. Can the real image of parcel distribution be obtained directly by using GUI？In ParaView, if it is not adjusted, parcels are displayed so large that they are overlapped together. Can you please give me some advice to get a relatively real image of parcel distribution in ParaView? Thank you.

In the GUI I have code that handles all the parcel scaling correctly. In Paraview, you need to change a couple settings:

- load the particle data (either the *.pvd file or the *.vtp stack).
- Select the
`Point Gaussian`

mapper from the drop down. - Change the
`Gaussian Radius`

to your particle radius.

Thank you, Weber. I noticed that the information of particles of different kinds is stored in the same file. When particle sizes are different, which value should the value of Gaussian Radius be set to?

Hi, Weber

Can you please tell me which version of ParaView you used? The option “scaling Mode” is not shown in my version.

I am using Paraview 5.9. I don’t know what the “scaling mode” actually does…

My recommendation to visualize particles is to:

- Always save the diameter in the vtp files
- Use point Gaussian representation
- Use a scale array, and set the Diameter as the Gaussian scale array
- Do not use a scale transfer function
- Set the Gaussian radius to 0.5. This is a multiplier of scale array, so we are basically saying that the radius is half the diameter.

I am not sure what the “scaling mode” does in Paraview 5.9, but it is related to Ray tracing so it has no effect with point Gaussian representation.