[roucho]

Hi Foamers,
It's been a week, now. Did anybody have any news from either SourceForge or Engys, concerning Helyx-OS?
Thanks!

[wyldckat]

Greetings to all!

I haven't gotten an answer from Daniel yet (as I mentioned in my previous post #179), but the Dynaflow website (who are partners with Engys in some projects, according to one of the pages I read) has the following note:

Originally Posted by http://www.dynaflow.com/software/helyx-os-download

The Helyx-OS download page is temporarily unavailable.
We are sorry for the inconvenience.

So whatever is going on, they are not yet at liberty to say.

Best regards,
Bruno

[chegdan]

@Rao_BuidlingDesign @Trilby Tim @roucho @wyldckat

Just to follow up on the discussion, HELYX-OS is back online with a new project page (http://engys.github.io/HELYX-OS/) and we have migrated from SourceForge to github. I posted an announcement earlier. Please use the new issue tracker if you would like to report an issue.

[Trilby Tim]

Cheers! Yes, I noticed it was back and have just downloaded and installed it, looks good!

[shaiashe]

Hi,
Im using HELYX OS v2.3.1, and when I start a new case, I get this msg in the terminal:
0 - pool-3-thread-1 - ERROR IOUtils - Error reading file /home/user/.HELYX-OS/tmp/tmp_1456908133/dictData/HELYX-OS/createCase.mapFieldsDict: File '/home/user/.HELYX-OS/tmp/tmp_1456908133/dictData/HELYX-OS/createCase.mapFieldsDict' does not exist
2 - pool-3-thread-1 - ERROR IOUtils - Error reading file /home/user/.HELYX-OS/tmp/tmp_1456908133/dictData/HELYX-OS/createCase.customNodeDict: File '/home/user/.HELYX-OS/tmp/tmp_1456908133/dictData/HELYX-OS/createCase.customNodeDict' does not exist


anyone know how to fix this?
thanks,

[chegdan]

Hi shaiashe,

What is your platform? How did you install HELYX-OS? It is looking for a folder in your home directory that it uses to hold temporary files/settings and cannot find the folder.

[Antech]

Hello.

I tried to use Helyx-OS with snappyHexMesh as an alternative to Ansys Meshing that we usually use for our industrial cases. Generated mesh is visually OK except for distored edges. I remember that I fixed it with surfaceFeatureExtract but I cannot find an option to run tis utility automatically from Helyx-OS GUI. I can edit the mesh script and add surfaceFeatureExtract but it requires it's dictionary to run. This leads to inconvenient mesh creation procedure (compared with Ansys Meshing that generates good meshes but not acceptable hexa). Is there an option to launch surfaceFeatureExtract correctly using Helyx-OS GUI or I need to tweak case files by hand anyway?
Thanks.

[ Moderator note: Moved from http://www.cfd-online.com/Forums/ope...penfoam-2.html ]

[chegdan]

Andrew,

No problem, HELYX-OS can extract feature edges like you have done in the past.

1. Within HELYX-OS, go to your surface and select it with the right mouse button.
2. Navigate and select Lines>Extract to bring up the extract feature lines dialogue.
3. Adjust the feature angle and hit apply to visualize the extracted lines. TIP: To see the lines better you can hide the surface in the tree to the left by unchecking the surface under geometry.
4. When you are happy with what has been extracted, hit save and then you will find the extracted line under the lined entry in the tree. There you can perform some distance refinement, including refinement directly on the lines i.e. refine at a distance of 0.

Good luck and I hope this helps. There is an ongoing discussion at

http://www.cfd-online.com/Forums/ope...pyhexmesh.html

so feel free to post questions there later if you are having any issues (this thread may be moved there too).

[ Moderator note: Moved from http://www.cfd-online.com/Forums/ope...penfoam-2.html ]

[Antech]

chegdan
Hello, thanks for you suggestion! I tried this and I can see lines. I press the Save button, the Wall_line object appears in the GUI tree. I save the project and try to build the mesh but an error occures:

fileName::stripInvalid() called for invalid fileName Wall_line.eMesh

Sorry, but, IMHO, there is a bug. I searched the case dir and found "Wall_line .eMesh" file with the trailing space in the name (obviously the space is illegal in this name so the mesher exits with error). OK, I deleted lines in GUI and recreated them. They was automatically named "Wall_line " (with trailing space). OK, I renamed them to "Test" and I see no spaces in the GUI edit box, but there visually even several trailing spaces in the tree item name . If I click on the "Lines" group in the tree the "Wall_line " or "Test " item name disappears at all. Is it some compatibility issue? Is this (line extract) feature widely used?

OS: CentOS 6.7 x64
HelyxOS 2.3.0
OpenFOAM 2.3.0

I added an attachment with screenshots.

I built the mesh despite of GUI behavior (lines element in the tree without name). Now the Wall_line.eMesh in triSurface dir has correct name but the mesh is still distored on edges (see my previous picture). I don't need a refinement on edges, I just need them to correspond to geometry (to be sharp). Should I switch some option?

[ Moderator note: Moved from http://www.cfd-online.com/Forums/ope...penfoam-2.html ]

[chegdan]

Antech,

I'll investigate further on the issue. Though a trailing space is in the name in the GUI, things are written correctly and I am able to get a mesh using the eMesh. I can see that it will delete the name if I click around a bit so I will report this behavior to the developer and see if we can improve it.

Concerning your issue of snapping, this could be many things. Not enough refinement is the most common issue, hence the use of the eMesh with a higher level at a distance of zero. Make sure that the Options>snapping>explicit snap feature is true to use the emesh files. Also you could increase the number of snap feature iterations inside of Options>snapping.

[Antech]

chegdan
Thanks for your answer.
Explicit snap feature is on. It was a Resolve Feature Angle that was set to 100 degree to cope with refinement at corners that is not needed and increase the mesh size. I set the resolve feature angle to default 30 degrees and voila - the edges are sharp but now they has undesired refinement I'm now playing with some sHM parameters in my spare time...

[chegdan]

Antech,

Ah yes, I assumed you didn't mess with the feature angle. Honestly the default settings work well for a majority of cases. There is a nice talk by Andrew Jackson on the different settings within SHM.

[MBttR]

Originally Posted by chegdan

Hi shaiashe,

What is your platform? How did you install HELYX-OS? It is looking for a folder in your home directory that it uses to hold temporary files/settings and cannot find the folder.

Hi Daniel,

I am not shaiashe but it seems you never got a reply. Well, here I am with the exact same problem!

I installed HELYX-OS according to the directions posted here. I am running OpenFOAM 2.4.0 on Ubuntu 15.10 on a VirtualBox machine on Windows 10 64 bit.

On the first boot of HELYX-OS, it asked for the core folder, here I gave:

/home/bruno/OpenFOAM/OpenFOAM-2.4.0

And ParaView Executable (of this I'm not sure if it is correct):

/home/bruno/OpenFOAM/ThirdParty-2.4.0/platforms/linux64Gcc/ParaView-4.1.0/bin/paraview

For this last one, there's also a ParaView folder in the ThirdParty folder so I was not sure.

Otherwise it seems to run fine, although it shows some behaviour which I'm not sure is normal;
I just tried opening the Cavity example from the OpenFOAM Tutorials, and it doesn't seem to read the existing blockMeshDict file. It just ignores this and creates a mesh according to whatever is dictated in HELXY-OS. I'm not sure if this is supposed to happen? The mesh I create in HELYX-OS does get read into ParaView later down the line though, so that's something.

Thanks!

Bruno

[chegdan]

Bruno,

Thanks for the report on your installation process. As for not reading in the constant/polyMesh/blockMeshDict file, HELYX OS expects that the blockMeshDict is in system. Most *Dict files are in the system directory (e.g. controlDict, snappyHexMeshDict, topoSetDict, etc.) but the blockMesh application has always looked in the constant/polymesh folder. HELYX OS reads/creates blockMeshDict files in the system directory (for consistency) and when blockMesh is run, it uses the command

blockMesh -dict ./system/blockMeshDict

to pass the correct dictionary in the system folder. Also, as for ParaView. Users do not need to point to the installation of ParaView in the thirdparty folder at all and can download ParaView (a newer version if you like) and just simply point HELYX-OS to that ParaView executable. This will easily allow users to update their install of ParaView if there is a new feature that you want to use.

Thanks for the report and please post if you have any more issues.

[MBttR]

Thanks Daniel! Will check it out as soon as I hit the office, but seems logical enough. Just wondering, how does HELYX-OS know it has to call blockMesh? Without a GUI, I tell the terminal to run blockMesh, but is it also stated somewhere in some system file that I was not yet aware of? Or does it do a random search for any file containing *Dict in the name? In case of the latter, what happens if there are two meshDicts available?

Cheers!

[chegdan]

Bruno,

Good question. It knows to call blockMesh since it must call it when you use snappyHexMesh. When you hit create, it calls blockMesh and then calls snappyHexMesh.

[MBttR]

Originally Posted by chegdan

Bruno,

Good question. It knows to call blockMesh since it must call it when you use snappyHexMesh. When you hit create, it calls blockMesh and then calls snappyHexMesh.

Hmm, my bad, I hadn't noticed HELYX-OS was for snappyHexMesh, should've read the title better Found the script where I could change that though.

Anyway, as expected it now loads the mesh fine when I move it to the system folder, but it ignores the names of the boundaries (hence movingWall, fixedWalls, frontAndBack) assigned to it in blockMeshDict. They get replaced by ffminx to ffmaxz. It magically read them correctly once, but I have no idea how I got that to happen.

It seems to load Runtime Controls input fine, with the correct inputs for startTime, endTime and intervals, but I think the Boundary Conditions are ignored as it gives No-Slip condition to everything. I also can't find my Initial Conditions, the Fields Initialisation tab is just empty, so I can't even see if my initial velocity and pressure are read correctly.

To summarize, keeping with the cavity tutorial, what would be the way to start from scratch and open this case in HELYX-OS? I thought, from scratch, just move blockMeshDict to the system folder and that's it, but I'm starting to think there's more to it?

Hope it all still makes sense. Thanks Daniel!

[MBttR]

Hi Daniel,

I was wondering if you have already found what the following error meant upon launching Helyx-OS? I have not used it since my last post, went back and dove into snappyHexMesh first as it seemed relevant to know the basis, but now I'm wanting to give Helyx-OS another try, as it seems helpful in trying to get a mesh to work properly.

It seems to launch OK but I'm not sure if I will run into fatal errors in the future. This is printed in the terminal:

0 - pool-3-thread-1 - ERROR IOUtils - Error reading file /home/user/.HELYX-OS/tmp/tmp_14566500559/dictData/HELYX-OS/createCase.mapFieldsDict: File '/home/user/.HELYX-OS/tmp/tmp_14566500559/dictData/HELYX-OS/createCase.mapFieldsDict' does not exist 
2 - pool-3-thread-1 - ERROR IOUtils - Error reading file /home/user/.HELYX-OS/tmp/tmp_14566500559/dictData/HELYX-OS/createCase.customNodeDict: File '/home/user/.HELYX-OS/tmp/tmp_14566500559/dictData/HELYX-OS/createCase.customNodeDict' does not exist

*edit*
Another question;
It seems Helyx-OS will only do MRF, but no rotating mesh, correct?

[chegdan]

if the GUI did not crash then do not worry about the warnings in the terminal. Make sure you have the most up to date version

It seems Helyx-OS will only do MRF, but no rotating mesh, correct

The next version that is coming out will have support for dynamic mesh. Those that attended OFW11 have a pre-release of HELYX-OS that will make its way to the github page in the coming weeks.

[djierian]

Originally Posted by chegdan

if the GUI did not crash then do not worry about the warnings in the terminal. Make sure you have the most up to date version

The next version that is coming out will have support for dynamic mesh. Those that attended OFW11 have a pre-release of HELYX-OS that will make its way to the github page in the coming weeks.

Is there an estimated release date you can provide? I am extremely interested in this feature! I'm sure many other people are as well.