Vectorial Helmholtz - Bent waveguide

General discussion about Elmer
kpereir
Posts: 7
Joined: 25 Mar 2021, 00:03
Antispam: Yes

Vectorial Helmholtz - Bent waveguide

Post by kpereir »

Dear all,

I am trying to replicate the tutorial for vectorial helmholtz, but with no luck. I am sure that I am missing something along the way, but couldn't figure out what. Really appreciate any inputs.

Attached is my .sif file and the log file. The Energy Functional value that I obtain is: -3037342.7152417749 504503.76872787910
I really don't know what I have done wrong.... I have already redone this tutorial a couple of times, but no luck in replicating the results.... :?

Thank you very much for any suggestions! :)
logfile.txt
(29.29 KiB) Downloaded 205 times
case.sif
(3.58 KiB) Downloaded 213 times
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Vectorial Helmholtz - Bent waveguide

Post by kevinarden »

It is not converging.
The tutorial has BiCGStabl your sif has BiCGStab
kpereir
Posts: 7
Joined: 25 Mar 2021, 00:03
Antispam: Yes

Re: Vectorial Helmholtz - Bent waveguide

Post by kpereir »

Thanks!!!! :)
kpereir
Posts: 7
Joined: 25 Mar 2021, 00:03
Antispam: Yes

Re: Vectorial Helmholtz - Bent waveguide

Post by kpereir »

Sorry for bringing up this question again. I am still struggling to get the same results as in the tutorial.... :cry: I am sure it is a silly mistake that I cannot figure out. May someone please help me?

I am currently using the BiCGStabl with vanka preconditioner to solve the Vector Helmholtz Equation. For the Vector Helmholtz Post Process, I am using BiCGStab (as the tutorial don't mention anything about changing it). Although, I have to say, I have also tried to change it, with no luck in replicating the results of Energy Functional value.

The tutorial says: Energy Functional value: -11284.937620324963 453999.53923919413
I keep getting: Energy Functional value: -2241.5418742151874 447371.37659609719

I am sure it is something really minor, but I cannot figure out. I would very much appreciate any comments on that, please. The sif file I am currently running is attached here.

Thanks again and sorry for the inconvenience.

Regards
Attachments
case.sif
with BiBGStabl solver
(3.58 KiB) Downloaded 188 times
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Vectorial Helmholtz - Bent waveguide

Post by kevinarden »

Not sure what to do with this in the tutorial since it is not an option in the GUI. meaning the inverse entry option.
Attachments
gui.png
(135.51 KiB) Not downloaded yet
Rich_B
Posts: 421
Joined: 24 Aug 2009, 20:18

Re: Vectorial Helmholtz - Bent waveguide

Post by Rich_B »

Interesting. As a test, I manually added that entry (Inverse Relative Permeability = 1) to the waveguide sif and this is the error message:

Code: Select all

MAIN: Reading Model: case.sif
LoadInputFile: Scanning input file: case.sif
LoadInputFile: Scanning only size info
LoadInputFile: First time visiting
LoadInputFile: Reading base load of sif file
ERROR:: LoadInputFile: 
ERROR:: LoadInputFile:  Unknown specifier: [1]
ERROR:: LoadInputFile:  In section: [material 1]
ERROR:: LoadInputFile:  For property name:[inverse relative permeability]
Also, that entry doesn't exist in Solver.Keywords, and doesn't show up in Models Manual.

Rich.
mika
Posts: 230
Joined: 15 Sep 2009, 07:44

Re: Vectorial Helmholtz - Bent waveguide

Post by mika »

Hi,

I think the vectorial Helmholtz solver doesn't read any keyword starting as "Inverse ...". Some earlier version perhaps did which may be the source of confusion. I believe the keywords described in Models Manual work.

-- Mika
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Vectorial Helmholtz - Bent waveguide

Post by kevinarden »

I went through the tutorial and my answer was the same as yours.
There may be some help in the test cases
https://github.com/ElmerCSC/elmerfem/tr ... zWaveguide
Rich_B
Posts: 421
Joined: 24 Aug 2009, 20:18

Re: Vectorial Helmholtz - Bent waveguide

Post by Rich_B »

To get back to the original question, here are a few comments on the sif file.

Either add 'Post File = case.vtu' in Simulation, or add the Result Output solver, but don't add both. In this case, I'd remove the Result Output solver.

To be sure that the vector helmholtz post process runs after vector helmholtz, change from
Exec Solver = Always
to
Exec Solver = After Simulation
in the vector helmholtz post process solver section.

Note that these two points are minor and probably don't have an impact on your issue.

Rich
Rich_B
Posts: 421
Joined: 24 Aug 2009, 20:18

Re: Vectorial Helmholtz - Bent waveguide

Post by Rich_B »

So I downloaded your second case.sif and added the elmer mesh.* files from the waveguide tutorial folder.

I commented out the 'Post File = case.vtu' line in Simulation and ran elmersolver.
VectorHelmholtzSolver: Energy Functional value: -11285.445827401698 453999.47783037461
The log.txt file is attached.

Rich.
Attachments
log.txt
(6.6 KiB) Downloaded 174 times
Post Reply