Subject: Re: Meshing problem Author: Tom Davis Date: 2002-07-23 16:55:00Thanks to Herbert, Bob and Terence for all your useful suggestions.
What I ended up doing was refining the mesh on the ajoining volume. This allowed me to mesh the second volume. I had to remove a number of b.c.'s from the first volume to do the refinement. None of this cleared up the pretension problem that I alluded to but did not describe in my original post. For some reason, the pretension elements were being generated outside of the plane that I had created for the purpose of generating them. I tried everything with no luck. I finally achieved a decent pretension mesh (I think) by moving the "cut" to a slightly up the bolt. I will now run the job to see if it worked. I'll know tomorrow morning.
Thanks again.
TWD
Thomas W. Davis, P. E. FSI International (http://www.fsi-intl.com/) 3455 Lyman Boulevard Mail Station 4-1035 Chaska, Minnesota 55318 PHONE 952-361-7359, FAX 952-361-7393 tom.davis@f...
I feel your pain. This is how I nurture a model back to good health. Something usually works.
1. If you can CDWR and CDRE it, sometimes that will fix it up.
2. There's a cool JAS anfout.mac in the archives and I think it's on Sheldon's website. That sometimes works when CDWR/CDRE fails. Seems to be quite robust. I use it a lot.
3. Sometimes, a VSEL, S, VOLU, , BadActor, , , 1 $ VDEL, BadActor $ VA, ALL will fix it up.
4. Change the ESIZ. Sometimes larger works, sometimes smaller.
5. Regen the areas that are connected to the lines that are connected to the offensive KP. Sometimes you have to do this with an ASUB (overlay option). Delete the original volume and areas and recreate the volume with the new one.
6. When I use pretension elements, I cut the model where I need to put them and then insert them with a macro. Seems to work better than letting ANSYS do it. Here's a code snippet of how I do it. I've never had much luck doing it in the GUI. Things always seem to get hosed up.
ALLS NUMC, ALL CSYS *GET, n, NODE, , NUM, MAX VSEL, S, VOLU, , 2, 3, 1, 1 pilot = n+1 N, pilot, , BoltCirc/2, FlngThk/2 PSME, 1, , pilot, ALL, , 0, Z, FlngThk/2, , , etPRETNS
I don't have a clue what happens but I've been there. I'm not really sure why these things work but they usually do. Usually a few hours of putzing around gets me back on track. For me, I've found that taking the few extra minutes to write a macro to insert pretension or gap elements is well worth it in the long run.