OpenParEM
which part of the code is used for data processing using the finite element method #5
Replies: 3 comments 2 replies
-
|
Is it a fem3D file? Are there other parts? |
Beta Was this translation helpful? Give feedback.
-
|
For post-processing for S-parameters, calculateS is called in OpenParEM3D.cpp, which can be found in port.cpp as BoundaryDatabase::calculateS. For saving fields for viewing with ParaView, save3DParaView is called in OpenParEM3D.cpp, which can be found in port.cpp as FieldSet::save3DParaView. |
Beta Was this translation helpful? Give feedback.
-
|
Why does the OPEM_L2ZZErrorEstimator function in the OpenParEM3D/src/OPEM_L2ZZErrorEstimator.cpp file use CurlCurlIntegrator to capture the left half of the wave equation in formula (5) in the OpenParEM3D/doc/OpenParEM3D_Theory_Methodology_Accuracy.pdf |
Beta Was this translation helpful? Give feedback.
-
|
The Zienkiewicz-Zhu error estimator calculates an error metric to identify mesh elements that are too coarse. It uses the physics of the problem being solved, and for electromagnetics, that is the wave equation. Both the physics equation and the flux are required, and MFEM has limited support for the flux calculations. The only flux calculation available related to the wave equation is for the curlcurl integrator, which is only half of the wave equation. It would be best to have the full wave equation, but that would require development of a custom MFEM method, which is very challenging. Using just the curlcurl part of the wave equation in the Zienkiewicz-Zhu error estimator is an approximation, but it seems to be effective in implementing adaptive mesh refinement. An improvement would be to add the linear portion of the wave equation to the error estimator with a custom method, and likely the adaptive mesh refinement would improve and enable smaller meshes for a given level of accuracy. So far, a test case has not be identified where the adaptive mesh refinement using just the curlcurl operator adaptive meshes in the wrong areas. |
Beta Was this translation helpful? Give feedback.
-
|
The reason is given in the last paragraph before section 4.1 on page 13 of the document “OpenParEM3D_Theory_Methodology_Accuracy.pdf”. The paragraph starts with “Note that AMR …”.
Not stated here though is that the EM field has to satisfy the wave equation in H as well as E. It is equally valid to work with H instead of E.
From: zzgbaifeng ***@***.***>
Sent: Monday, December 16, 2024 2:26 AM
To: OpenParEM/OpenParEM3D ***@***.***>
Cc: Brian Young ***@***.***>; Comment ***@***.***>
Subject: Re: [OpenParEM/OpenParEM3D] which part of the code is used for data processing using the finite element method (Discussion #5)
Thank you very much for your explanation, but now I still have a question: why is the grid function of H brought into the curlcurl integrator, while the corresponding grid function in the wave equation should be E?
—
Reply to this email directly, view it on GitHub <#5 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AVSFWC6QWUHRODGJZDTNXLL2F2E7VAVCNFSM6AAAAABR22ARNWVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCNJXHAYTINQ> .
You are receiving this because you commented. <https://github.com/notifications/beacon/AVSFWC2KZBCP4LZHAQKEG6L2F2E7VA5CNFSM6AAAAABR22ARNWWGG33NNVSW45C7OR4XAZNRIRUXGY3VONZWS33OINXW23LFNZ2KUY3PNVWWK3TUL5UWJTQAWCVSE.gif> Message ID: ***@***.*** ***@***.***> >
|
Beta Was this translation helpful? Give feedback.
-
Could you please tell me which part of the code is used for data processing using the finite element method? Thank you very much.
Beta Was this translation helpful? Give feedback.
All reactions