[Gmsh] Harmonic To Time error
paulz at ieee.org
Tue Nov 28 06:17:23 CET 2017
I have what thatappears to be computational error in the HarmonicToTime plug-in. This happened while using onelab to demonstrate EM fields inwaveguides. Set up:
- In gmsh, open models/waveguides/main.pro
- In the dialog for Input > Problem, use the pull-down menu and select Rectangular waveguide [3D].
- Under Post-processing, select  e. This is the electric field.
- Adjust the view to see the long (Y) axis. On any of the long sides, you should see 3 places with a maximum E field alternating with 3 minimums of the E field.
- Use the Step Forward from the animation icons in the bottom left border. The display alternates between “e - real part” and “e – imaginary part”. The locations of field maximums change; amplitudes stay the same. The scale is 0 to 502.
Now run theHarmonicToTime plug-in with the default settings. The step functioncauses the peaks to move in smaller increments, as expected. Howeverthe amplitude changes during the cycle. It is maximum at step 9 andminimum at step 0. And the scale has changed to be 0 to 542.
Further experimentswere done with various numbers of steps. With NumSteps =4, everythingappears as expected. Full scale is 502. Step 0 and Step 1 lookidentical to the real and imaginary parts of e, as is expected. Steps 2 and 3 continue the cycle with the same amplitude.
Minor problem: thedefault time step is -1, aka going back in time.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the gmsh