[hsezsz]

Hallo Everyone,

I am doing simulation of free burning arc with CFX. I meet a prolem with the Electric Potential Equation Residual. I set my convergence criterion to 1e-4. At first all the residuals are quite good and the adaptive time steps keep increasing. But after some time, all the other residuals are quite ok far below 1e-4, but only the Electric Potential Equation Residual is above 1e-4 which makes the adaptive time step doesn't work to increase my time step and stay constant at 1e-7s. It is too slow for my simulation. Could anyone help me how to solve this problem? Thank you very much!

The out file data is as follows

COEFFICIENT LOOP ITERATION = 20 CPU SECONDS = 1.722E+06
----------------------------------------------------------------------
| Equation | Rate | RMS Res | Max Res | Linear Solution |
+----------------------+------+---------+---------+------------------+
| U-Mom | 0.99 | 1.6E-10 | 6.7E-08 | 8.1E-04 OK|
| V-Mom | 1.00 | 1.9E-10 | 1.1E-07 | 8.0E-04 OK|
| W-Mom | 0.97 | 8.8E-11 | 4.7E-08 | 8.0E-04 OK|
| P-Mass | 1.00 | 1.1E-09 | 6.3E-07 | 4.9 1.1E-04 OK|
+----------------------+------+---------+---------+------------------+
| Ax-Mag-Air | 0.39 | 5.7E-17 | 8.4E-10 | 7.7E-02 OK|
| Ay-Mag-Air | 0.35 | 1.6E-17 | 3.9E-10 | 7.9E-02 OK|
| Az-Mag-Air | 0.36 | 1.2E-16 | 1.0E-09 | 9.6E-02 OK|
| Ax-Mag-EleB | 0.49 | 2.2E-18 | 5.8E-10 | 7.7E-02 OK|
| Ay-Mag-EleB | 0.57 | 9.2E-18 | 1.7E-09 | 7.9E-02 OK|
| Az-Mag-EleB | 0.59 | 1.4E-17 | 2.2E-09 | 9.6E-02 OK|
| Ax-Mag-EleT | 0.78 | 1.9E-11 | 1.7E-09 | 7.7E-02 OK|
| Ay-Mag-EleT | 0.78 | 1.4E-11 | 1.1E-09 | 7.9E-02 OK|
| Az-Mag-EleT | 0.79 | 2.3E-11 | 2.1E-09 | 39.0 9.6E-02 OK|
+----------------------+------+---------+---------+------------------+
| E-Pot-Air | 1.05 | 5.7E-04 | 1.7E-01 | 8.2E-02 OK|
| E-Pot-EleB | 0.78 | 5.6E-11 | 5.2E-09 | 8.2E-02 OK|
| E-Pot-EleT | 0.79 | 2.0E-16 | 1.4E-14 | 14.3 8.2E-02 OK|
+----------------------+------+---------+---------+------------------+
| H-Energy | 0.99 | 7.3E-11 | 4.6E-08 | 6.0E-09 OK|
| T-Energy-EleB | 0.50 | 1.4E-11 | 4.5E-10 | 6.0E-09 OK|
| T-Energy-EleT | 0.50 | 2.3E-11 | 8.7E-10 | 5.6 6.0E-09 OK|
+----------------------+------+---------+---------+------------------+
| K-TurbKE | 0.98 | 2.1E-11 | 7.8E-09 | 5.6 1.8E-09 OK|
| E-Diss.K | 0.97 | 1.3E-11 | 4.6E-09 | 10.6 9.7E-13 OK|
+----------------------+------+---------+---------+------------------+


================================================== ====================
| Adaptive Timestepping Information |
----------------------------------------------------------------------
| Direction | Ratio | Last Value | Next Value | RMS Co | Max Co |
+----------------+-------+------------+------------+--------+--------+
| Unchanged | 1.000 | 1.0000E-07 | 1.0000E-07 | 0.02 | 0.22 |
+----------------+-------+------------+------------+--------+--------+

[Opaque]

You can create a backup file including the residuals for all the equations.

Open the backup in CFD-Post and create a monitor point for the location of the Variable Maximum, Select the residual of interest, say E-Pot.Residual and you will find out where the equation is refusing to converge.

Once you have the location, you can study mesh quality in the such region, solution characteristics, etc and perhaps diagnose what might be wrong with your case.

Hope the above helps,

[hsezsz]

Originally Posted by Opaque

You can create a backup file including the residuals for all the equations.

Open the backup in CFD-Post and create a monitor point for the location of the Variable Maximum, Select the residual of interest, say E-Pot.Residual and you will find out where the equation is refusing to converge.

Once you have the location, you can study mesh quality in the such region, solution characteristics, etc and perhaps diagnose what might be wrong with your case.

Hope the above helps,

Hi Opaque,

Thank you so much, your method works. The reason for the problem is the mesh density near the electrodes are too low. If I increase the density, it disappears!

For the people who might be interested in this problem, I use the plane to show the residual which is easier to locate the problem than point. But this is method is definitely useful.

Thank you very much!