Electric conductivity is outside of the Argon range

[rodion]

Hello everyone,

I am trying to build a simulation of electric arc in Argon with imposed electric current density at the cathode surface instead of imposed electric potential difference.
Hence, the boundary conditions are as following:
Anode: Pr=0.0
Cathode: grad(Pr) = |-J(imposed surface distribution)/Sigma(from the adjacent cell)| > 0
All other surfaces: grad(Pr) = 0.0
The chosen solver for electric potential is GMRES.
I am using Code_Saturne 5.0.5 on Windows.

But I receive very strange results. For example:
1. there are small spots with abnormally high positive electric potential, while in the whole domain it should be lower than the value at the anode. Near this small spots with high potential very high electric current appears and the arc crashes. Also the whole electric potential distribution fluctuates a lot and changes the shape from step to step.
2. in some cells electric conductivity is way higher than the max value or lower than the min value of the provided data for Argon. I.e. the range in dp_ELE is [0.13214E-03;12300.], while in the output data I get electrical conductivity in the range [9.5527e-006;4.9453e+006]. These values cannot be obtained by extrapolation to the corresponding enthalpy values (from 0.22037E+09 in dp_ELE to 3.5835e+010 in output). Also I have clipping of enthalpy at the min value.

Does somebody know, what can be the reason of such Code_Saturne behavior?
Am I doing something wrong?

Best regards,
Rodion

[Yvan Fournier]

Hello,

This may be due to numerical sensitivity or mesh sensitivity and values in computed scalar variables not respecting the "min-max" property. In the latter case, defining max values matching the boundary conditions for the enthalpy and/or electric potential might be an (ugly but workable) solution. The GUI does not seem to allow defining clippings for the electric potential (for no good reason, probably an omission), so that must be defined with user subroutines.

I have not checked in detail how a prescribed electric current is translated in the code, so there might be an issue there. I'll try to check with an ex-developer who has worked on electric arcs before.

How "clean" is your mesh ? Do you notice anything else in the "listing" file ?

Regards,

Yvan

[rodion]

Thank you for your answer, Yvan.

Could you please tell me, what should I specify in the Electrical models/Electric arc/Variables scaling model?
If I want to impose electric current as a boundary condition, I do not need any scaling here, right?
But if I just choose "user defined", the electric potential and electric current fields become very low and almost constant.
As far as I can see the variable "pot_diff" affects on the electric potential field even if pot_diff is not used in boundary conditions.

In the listing i get the following information about the mesh:

Mesh
Number of cells: 250560
Number of interior faces: 745660
Number of boundary faces: 12040
Number of vertices: 256741

Groups:
"0"
cells: 250560
interior faces: 3020
boundary faces: 12040
"ANODE_BC"
boundary faces: 2200
"ANODE_SH"
interior faces: 2240
"ANODE_SHEATH"
cells: 2200
"ANODE_SIDE_BC"
boundary faces: 400
"CATHODE_BC"
boundary faces: 700
"CATHODE_FLUID"
boundary faces: 1280
"CATHODE_SH"
interior faces: 780
"CATHODE_SHEATH"
cells: 1400
"FLUID"
cells: 246960
"FLUID_BC"
boundary faces: 3560
"INLET"
boundary faces: 800
"OUTLET"
boundary faces: 3100

--- Information on the volumes
Minimum control volume = 1.3264092e-013
Maximum control volume = 3.3315884e-008
Total volume for the domain = 2.6970152e-005

Criterion 1: Orthogonality:
Number of bad cells detected: 0 --> 0 %

Criterion 2: Offset:
Number of bad cells detected: 108 --> 0 %

Criterion 3: Least-Squares Gradient Quality:
Number of bad cells detected: 0 --> 0 %

Criterion 4: Cells Volume Ratio:
Number of bad cells detected: 0 --> 0 %

Criterion 5: Guilt by Association:
Number of bad cells detected: 0 --> 0 %

Warning:
--------
Mesh quality issue has been detected

The mesh should be re-considered using the listed criteria.

The calculation will run but the solution quality may be degraded...

Computing geometric quantities (0.0791 s)

ALMAX = 0.29989E-01 (Characteristic length )

Best regards,
Rodion

[Yvan Fournier]

Hello,

I know from discussions with a developer of the electric arcs module that computations can be quite sensitive to how scaling is done, and there is need for improvements there, but am otherwise not knowledgeable enough in this module to provide more useful help. I'll recommend that person (who now moved to another job in our company) to check the forum if he has time.

Regards,

Yvan

[rodion]

Thanks a lot, Yvan.

I found the problem. I have imposed the value of electric potential gradient instead of the scalar product of the electric current vector and the normal vector. I just had to read the user guide more carefully.
Also the idea to choose "user defined" for scaling was correct. With "user defined" electric potential field becomes stable and reasonable, while pot_diff is just constant.

Best regards,
Rodion