Linear profile even with fine mesh

[JaeHyun-Nam]

Hello all,

I made fine mesh using Gmsh like below with special attention to gas-liquid interface. The minimum size of is ~5x10^-10m.

However, below results have discontinuous profiles near the interface and seem to have very coarse mesh.

(but in gas phase, profile is very continuous and smooth)

How can I explain this?

P.S. I used Paraview that opened exodus format file. Is there any possibility that this format file contains nodes-skipped values?

Thank you!

[csdechant]

Do you have the .geo file for the mesh you made in a Git branch? It is a little hard to tell which domain is the gas and which is the liquid from the picture. My first guess is that there was a refinement issue and only the gas domain got refined. As for Paraview, there have been know issues with line interpolation between nodes (usually with 2-D simulations) , but I haven't heard about Paraview skipping nodes.

[JaeHyun-Nam]

Hello, this is .geo file.

liquidNew_modified.txt

This is file is from "liquidNew.geo" I made some modifications here to refine liquid domain.

[csdechant]

I ran the modified mesh with mean_en.i in the /tests/1d_dc/ directory. The results for the electrons in the water are shown in the plot for both mesh versions. The mesh refinement seems to be good. Are you also running a modified version of mean_en.i?

[Jae-Nam]

@csdechant Thank you. I also just ran refined mesh with "mean_en.i" and result seems to be good.
But with my modified input file (here I added some reactions using crane action and scaled up liquid domain 1e-7=>1e-5 m), there are still discontinuous profiles. In this case, can the problem be convergence related?

[Jae-Nam]

I figured out. This was just a 'resolution' problem of Paraview. I changed to higher resolution settings and the profile became more continuous.