Learning Forum

[emh711]

I did a sweep.

[Taylor Robertson]

Hello
The reason for this is likely that the matrix size of the fields as changing as you increase the simulation span. The simulations have run, and they have results, it is just that when the sweep tries to create a dataset of the resulting E, H field matrices by stitching them together there is a dimensional mismatch. If you reload the results you should see an error.
https://support.lumerical.com/hc/en-us/articles/360034922933
I would make the monitors smaller than the simulation span, and ensure the mesh is consistent between simulations even as you add mesh cells at the boundary. If the results, have the same geometric extent, and mesh cell number there should be no issue. The transmission is measured across the entire monitor, and will therefore not be sensitive this grid mismatch issue.

Best

[emh711]

I tried the method that you mentioned.
I made the monitor smaller than the simulation span.
I am still unable to obtain the sweep results of E and M fields.
I am not sure how I can make sure the mesh is consistent between simulations since the x span and y span of FDTD simulation span change between simulations.
That would have caused inconsistency of the mesh.
What is your suggestion of what I should do next to fix this problem?

[sagar]

I am also facing the same issue. I increased the simulation time and disabled the autoshutoff also but still did not work. I could not understand the actual reason.

[sagar]

I am facing the same issue. After putting mesh around the simulated object, during the sweep of the X_span, I could not able to produce the electric field profile it is producing transmission value. But without mesh, it is able to produce both electric field and transmission coefficient.

[Taylor Robertson]

If you are sweeping the span of FDTD, then the sizes of the matrices will differ, nothing you can do in that case. Measuring T is fine, because this is a scaler, but the matrices must be the same size to be packaged into a single dataset. I didn't entirely understand your intention, and to do this you would need to mesh to remain constant and the span to be the same.
Perhaps you could add some code to re-interpolate onto a predefined grid? So you would take the fields, and use https://support.lumerical.com/hc/en-us/articles/360034925893-interp-Script-command on a grid the same size as the largest span you care about. The for each simulations interpolate it onto this grid. Likely setting the fields outside the simulated span, for the smaller spans, to zero.
You could do this in the model analysis, and then instead of of getting the results directly from the monitor get the as results from the the analysis group and they would be guaranteed to be the same size.
https://support.lumerical.com/hc/en-us/articles/360045465813-Analyzing-Simulation-Results-Analysis-Groups-Introduction-to-Analysis-Groups