|
[Sponsors] |
July 5, 2005, 12:08 |
What's the point of imprting an stl mesh?
|
#1 |
Guest
Posts: n/a
|
When i import my stl file, I am able to mesh it and get it off t fluent but I can't set my boundary conditions. I end up importing only 1 face, (the surface of my object) and i can't set an in and out condition with just one face. Do i need to play around with the feature angle or what? This has been a long frustating problem, thanks.
|
|
July 5, 2005, 15:50 |
Re: What's the point of imprting an stl mesh?
|
#2 |
Guest
Posts: n/a
|
Set your boundary conditions in Gambit...
Under solver, make sure you pick Fluent 5/6 Then you define the BCs with the third button under "Operation" in the top right portion of the screen... it looks like a green cube with a couple little cubes and pointers. Don't worry about setting your continuum entity (under "Zones", that's the all green cube), because gambit will give it a default Fluid continuum. Define all of your face BCs though (unders "Zones", that's the button where the cube only has one face highlighted in green). If you need more faces to define BCs to (say your imported stl file came in as only one face for whatever reason... I don't even know if that's possible, but we'll just assume it is for the moment), then you can always split the existing faces in Gambit before you mesh your model. Hope this helps, and good luck, Jason |
|
July 5, 2005, 17:52 |
Re: What's the point of imprting an stl mesh?
|
#3 |
Guest
Posts: n/a
|
well the original problem was i couldnt split the faces and set the boundary conditions in gambit. however i was able to import with the feature angle at 180 into gambit . It took longer, but gave me faces that i could set boundary conditions off of.
Thanks though. You wouldnt happen to know if gambit works in parallel processing? I know fluent does. It's just that my gambit processing takes longer than my fluent processing. thanks |
|
July 6, 2005, 10:07 |
Re: What's the point of imprting an stl mesh?
|
#4 |
Guest
Posts: n/a
|
Gambit doesn't run in parallel (at least not that I've ever seen or heard of). Your Gambit processing takes that long? What are you doing in Gambit? Are you using sizing functions? If so, it helps a lot to split the volume, so that the sizing functions only have to be calculated over a small portion of the geometry. And using a smaller volume, you can lower the maximum size limit for that sizing function, which will also help speed up the processing. Also, curvature sizing functions take longer to process than fixed sizing functions. If you're running Gambit 2.2, then you have access to the meshed sizing function. This takes the longest to process, but it's very useful so it's hard to avoid using sometimes.
Hope this helps Jason |
|
July 6, 2005, 10:21 |
Re: What's the point of imprting an stl mesh?
|
#5 |
Guest
Posts: n/a
|
well when i import the stl with feature angle of 180, i enede up with 15k vertices, 35k edges, and 20k faces and 1 volume. So yea its pretty big, takes a while to load, and a long time to do anything. Im trying to set up the volume mesh, the boundry conditions, than seend it to processing thru fluent, how could i do that and have the volume split up into pieces? Also what is a sizing function, and how would i be able to change it?
|
|
July 6, 2005, 10:46 |
Re: What's the point of imprting an stl mesh?
|
#6 |
Guest
Posts: n/a
|
Take stl mesh in TGrid and mesh there. U can define BCs in Fluent. No need to go to Gambit.
|
|
July 6, 2005, 11:08 |
Re: What's the point of imprting an stl mesh?
|
#7 |
Guest
Posts: n/a
|
Actually... a quick side note... have you tried using a step file (.stp) instead of an stl file? I don't know how complicated your geometry is, or where you get it from, but I've had goodluck with step files coming through with the appropriate faces, and not having to assign any feature angles.
20k faces is excessive... that's just going to be a lot to deal with in Gambit, so I can defenitely understand why you're being dragged down. As far as splitting the geometry... you can create faces or volumes that represent where you want to split the main volume. Then use the split command with the connected option on. The connected option means there will be only one face shared between the two volumes. The mesh on this face will be shared by both volumes as well. When you define your BCs, don't define any BC to this face. When Gambit exports the mesh, it will ignore these interior faces so that you'll have a continuous mesh. If you do want them to show up in Fluent (say you know in advance you're going to want the velocity vectors on this plane) then you can define it as an "interior" BC. This is similar to creating a surface within Fluent... it's really only good for post-processing, and doesn't affect the flow at all. Splitting the volumes is a technique to help with meshing. Another thing to help with meshing is to use sizing functions (under the toolbox, it's the yellow grid that looks like a radar screen or something). It's a way of controlling the mesh. Using a sizing function you can define a "start size" and "growth rate" and a "size limit". You also define what geometry to start from (vertex, edge, face, or volume), and what geometry to act on (edge, face, or volume). So what happens, is the mesh will have the "start size" at the geometry you defined, and will grow at the "growth rate" you defined until it reaches the "size limit" or the end of the geometry that the sizing function acts on. The unstructured meshing algorithm in Gambit is notorious for being difficult to control the growth of the cells within a volume, but the sizing functions are the tool to fix that. The sizing function I described before is the "fixed" sizing function. There's also a "curvature" sizing function where instead of a "start size" you define an angle that represents the minimum angle between face normals. Also, there's a "meshed" sizing function. You mesh a face with whatever mesh you want, then the "meshed" sizing function will grow the mesh from this face at the growth rate you define up to the size limit you define. Fluent has some good tutorials on sizing functions and decomposing geometry (that's splitting faces and volumes to make meshing easier). A note on growth rates... I don't recommend a growth rate over 1.2, and you should try to stay down around 1.1 to 1.2 in areas of complicated flow. If your mesh grows too fast it can distort the flow. If you search the forum (search link all the way at the top of the page, make sure you're looking in the Fluent forum) there's a lot of good information on sizing functions here as well. Hope this helps, and good luck, Jason |
|
July 6, 2005, 15:37 |
Re: What's the point of imprting an stl mesh?
|
#8 |
Guest
Posts: n/a
|
In regards to tgrid, ive tried that route, with stl, after it gets to fluuent i cant set the boundrary conditions because i would only have the surface face rather than any individual faces. thanks though
in regards to Jason's message Thanks, I have tried step, and with the new way my stl files are created it might be easier to do than previously( to set the face boundary conditions) however i end up with step files 40mb + from a 2mb stl file (going thru rhino) unless you know a way to write a step file directly from idl. also the step file doesn't seem to import into gambit with the face meshed, so that is a time consuming step that i can skip with the stl. I appreciate it though, you gave me some good ideas that should take me a while to try. Thanks |
|
July 6, 2005, 17:51 |
Re: What's the point of imprting an stl mesh?
|
#9 |
Guest
Posts: n/a
|
oops, acidently posted a message before yours, check tha tout if you can. I'm using idl to write polygon and vertices of the surface (its a complicated surface, material microstrucure) to stl file, since its the only format i have found to write the data to directly. if you know of a way to write step file directly from idl, let me know, thanks a lot
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
[snappyHexMesh] Number of cells in mesh don't match with size of cellLevel | colinB | OpenFOAM Meshing & Mesh Conversion | 14 | December 12, 2018 09:07 |
Moving mesh | Niklas Wikstrom (Wikstrom) | OpenFOAM Running, Solving & CFD | 122 | June 15, 2014 07:20 |
3D Hybrid Mesh Errors | DarrenC | ANSYS Meshing & Geometry | 11 | August 5, 2013 07:42 |
Icemcfd 11: Loss of mesh from surface mesh option? | Joe | CFX | 2 | March 26, 2007 19:10 |
CFX4.3 -build analysis form | Chie Min | CFX | 5 | July 13, 2001 00:19 |