[ICEM] Wing Woes. Difficult geometry,(or inept user) Swept and tapered 3d wing

ShowponyStuart · May 12, 2013, 02:21

I already did an o-grid around this wing (like in the vid)? Plus I have deleted the blocks inside the wing.

Far · May 12, 2013, 02:24

When you select the blocks inside the wind for o-grid generation, do not select any face.

ShowponyStuart · May 12, 2013, 02:35

I will give that a go, but I am still trying to figure out why I got those weird results I mention earlier.

There is nothing i this area so i figure it has to be an artifact of the mesh. And its in one of the big blocks in the farfield, I would have been less surprised if it had happened where there was heaps of blocks (inline with the wing but towards the wall opposite the one the wing is attatched too)

So yeah, im real confused.

Far · May 12, 2013, 02:48

Quote:

Originally Posted by ShowponyStuart

I will give that a go, but I am still trying to figure out why I got those weird results I mention earlier.

There is nothing i this area so i figure it has to be an artifact of the mesh. And its in one of the big blocks in the farfield, I would have been less surprised if it had happened where there was heaps of blocks (inline with the wing but towards the wall opposite the one the wing is attatched too)

So yeah, im real confused.

Observing flow field without number (cl, cd or whatever variable) will confuse you. Before getting a decent simulation (acceptable mesh, acceptable parameters and convergence) observing merely flow field will mislead you.

At the moment you task should be to get a good mesh. Even after getting numbers I do not look on flow field unless necessary to find reason for some change in data set.

ShowponyStuart · May 12, 2013, 02:53

I did monitor my cl and cd and while they were roughly in the region I expected them they were not converging as nicely (werent particularly stable)

But the reason I pointed that out, is that there is something wrong with my mesh in the region I thought was good quality. Its almost like there is some solid object or something in that region (looking at streamlines shows them moving around "something"). So im not sure if there are some solid elements that I generated in my mesh by accident or something

So I agree with you completely, Im just trying to figure out why the mesh quality is poor in those areas.

Far · May 12, 2013, 02:59

wrong mesh in one area may give you problem area in other region as numerical error is accumulating (numerical diffusion etc) due to wrong calculation of flux, thats why I saying do not monitor flow field at the moment. It does not mean you never look on flow field

.

Once your mesh is good, most of the problems will go away, believe me

. Just put hard work on mesh, as it is the first step to successful simulation.

ShowponyStuart · May 12, 2013, 05:15

Again I cant fault your logic, and I agree with you completely.

I just get a bit flighty sometimes when things arent going right, even if I know im doing the right thing haha.

I think Ive made a small breakthrough with the blocking (given your pointer before.) With any luck I will be back with good news soon.

Far · May 12, 2013, 08:13

Quote:

Originally Posted by ShowponyStuart

Again I cant fault your logic, and I agree with you completely.

I just get a bit flighty sometimes when things arent going right, even if I know im doing the right thing haha.

I think Ive made a small breakthrough with the blocking (given your pointer before.) With any luck I will be back with good news soon.

After several years in CFD field (that does not make me expert to give advises

but I still do

) once your mesh is getting right, your simulation work will make progress at very high pace. During meshing time is flying and one is not making any progress (apparently). Another advantage is that, this experience will help you in next projects, so keep it up and I hope you will get things done.

In meshing (specially blocking) the approach should always be to : one step back and two steps forward instead One step forward, two steps back. (I am in no way a philosopher

). So abandon old blocking and start from scratch and start to build up things and avoid previous mistakes. Also keep in mind the blocking I have attached (DWP2 made by Simon)

Far · May 12, 2013, 08:51

Just take a look at following threads. It took him around two months or so to get good blocking and mesh (even more as his case very highly complicated but he done it). Really impressive work.

http://www.cfd-online.com/Forums/ans...w-winglet.html

http://www.cfd-online.com/Forums/ans...w-winglet.html

ShowponyStuart · May 12, 2013, 22:01

Havent had a chance to read through those links yet (will do that now) but this is my latest attempt. I think that while the blocking is getting closer to what it should be, there are still a lot of problems.

My latest files too.
https://www.dropbox.com/s/kyykrrqayg...4AOA_mach4.zip

ShowponyStuart · May 13, 2013, 06:28

Arghhh...Im about ready to give up.

It was all going good...Reasonably nice mesh you say?

..hey that blocking looks reasonably tidy, you might get a mesh with reasonable quality...

Well if you thought that you would be wrong (as was I).

Just thought I was starting to wrap my head around it a bit, but apparently not.
Im going to have to call it quits soon. I could be working on this until after the project is due for all I know. Unless I am hit with some epiphany, or some assurance I can reliably build the mesh quickly by the end of the week, I dont want to waste any more time chasing my tail.

My last set of files for anyone interested
https://www.dropbox.com/s/28fmx3oyyq...OA_mach5_2.zip

ShowponyStuart · May 13, 2013, 07:53

Is it possible that it is some sort of connectivity issue or something? Or possible due to the way I created the O-grid. (Just selected the wing block and no faces like far suggested.) Perhaps I needed to select the face where the wing connects to the wall?

ShowponyStuart · May 15, 2013, 03:52

Does anyone have an idea of what is going wrong here? Ive been doing some googling and stuff and still cant quite figure out what is wrong. I have the impression that its only a small problem, and if I can get it fixed, my mesh should come together nicely.

It is telling me that the highlighted block is the worst block, though perhaps it is a couple of the outlet blocks that is causing an issue.

This is what happens with the outlet mesh. Perhaps some sort of connectivity or association issue??

And there doesnt seem to be a problem with any of the other sides, just when I select the outlet.

So does anyone have any thoughts on this, im pretty stuck and could use some input.

https://www.dropbox.com/s/28fmx3oyyq...OA_mach5_2.zip

FJSJ · May 15, 2013, 07:18

I´ve just used the "associate edge to surface" tool and "associate face to surface" tool to resolve your problem in your outlet part.

However, I´ve checked your quality determinant 3x3x3 and I think I´ve seen bad elements...

https://www.dropbox.com/s/63go4kbobm...OA_mach5_2.rar

ShowponyStuart · May 20, 2013, 04:58

I missed this that last post somehow but thanks for that FJSJ. It seemed to work a treat.

I played around a bit and got the quality up. I got pretty much all the elements above 0.9 for the 3x3x3 determinate, though there are are couple of sketchy elements at the trailing edge mostly (and a handful at the wing tip)

Problem is now I am getting this weird error when I go to import my mesh into CFX. And it wont load up any of the domains or anything. I have tried renumbering the mesh and creating new parts in attempt to fix the issue but I havent been sucessful yet.

Im so close to being able to run a sim that I can taste it, but have been beaten again! Does anyone have any ideas?

latest files:
https://www.dropbox.com/s/c972geny1f...oa_mach5_5.zip

ShowponyStuart · May 20, 2013, 20:06

No one has any thoughts on how to fix this error?

May 12, 2013, 05:15		#27
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	Again I cant fault your logic, and I agree with you completely. I just get a bit flighty sometimes when things arent going right, even if I know im doing the right thing haha. I think Ive made a small breakthrough with the blocking (given your pointer before.) With any luck I will be back with good news soon. Far likes this.

May 12, 2013, 22:01		#30
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	Havent had a chance to read through those links yet (will do that now) but this is my latest attempt. I think that while the blocking is getting closer to what it should be, there are still a lot of problems. My latest files too. https://www.dropbox.com/s/kyykrrqayg...4AOA_mach4.zip Far and blgypeng like this.

May 13, 2013, 06:28		#31
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	Arghhh...Im about ready to give up. It was all going good...Reasonably nice mesh you say? ..hey that blocking looks reasonably tidy, you might get a mesh with reasonable quality... Well if you thought that you would be wrong (as was I). Just thought I was starting to wrap my head around it a bit, but apparently not. Im going to have to call it quits soon. I could be working on this until after the project is due for all I know. Unless I am hit with some epiphany, or some assurance I can reliably build the mesh quickly by the end of the week, I dont want to waste any more time chasing my tail. My last set of files for anyone interested https://www.dropbox.com/s/28fmx3oyyq...OA_mach5_2.zip blgypeng likes this.

May 12, 2013, 02:21		#21
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	I already did an o-grid around this wing (like in the vid)? Plus I have deleted the blocks inside the wing.

May 12, 2013, 02:24		#22
Far Super Moderator Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,553 Blog Entries: 6 Rep Power: 54	When you select the blocks inside the wind for o-grid generation, do not select any face.

May 12, 2013, 02:35		#23
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	I will give that a go, but I am still trying to figure out why I got those weird results I mention earlier. There is nothing i this area so i figure it has to be an artifact of the mesh. And its in one of the big blocks in the farfield, I would have been less surprised if it had happened where there was heaps of blocks (inline with the wing but towards the wall opposite the one the wing is attatched too) So yeah, im real confused.

May 12, 2013, 02:53		#25
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	I did monitor my cl and cd and while they were roughly in the region I expected them they were not converging as nicely (werent particularly stable) But the reason I pointed that out, is that there is something wrong with my mesh in the region I thought was good quality. Its almost like there is some solid object or something in that region (looking at streamlines shows them moving around "something"). So im not sure if there are some solid elements that I generated in my mesh by accident or something So I agree with you completely, Im just trying to figure out why the mesh quality is poor in those areas.

May 12, 2013, 02:59		#26
Far Super Moderator Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,553 Blog Entries: 6 Rep Power: 54	wrong mesh in one area may give you problem area in other region as numerical error is accumulating (numerical diffusion etc) due to wrong calculation of flux, thats why I saying do not monitor flow field at the moment. It does not mean you never look on flow field . Once your mesh is good, most of the problems will go away, believe me . Just put hard work on mesh, as it is the first step to successful simulation.

May 12, 2013, 08:51		#29
Far Super Moderator Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,553 Blog Entries: 6 Rep Power: 54	Just take a look at following threads. It took him around two months or so to get good blocking and mesh (even more as his case very highly complicated but he done it). Really impressive work. http://www.cfd-online.com/Forums/ans...w-winglet.html http://www.cfd-online.com/Forums/ans...w-winglet.html

May 13, 2013, 07:53		#32
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	Is it possible that it is some sort of connectivity issue or something? Or possible due to the way I created the O-grid. (Just selected the wing block and no faces like far suggested.) Perhaps I needed to select the face where the wing connects to the wall?

May 15, 2013, 03:52		#33
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	Does anyone have an idea of what is going wrong here? Ive been doing some googling and stuff and still cant quite figure out what is wrong. I have the impression that its only a small problem, and if I can get it fixed, my mesh should come together nicely. It is telling me that the highlighted block is the worst block, though perhaps it is a couple of the outlet blocks that is causing an issue. This is what happens with the outlet mesh. Perhaps some sort of connectivity or association issue?? And there doesnt seem to be a problem with any of the other sides, just when I select the outlet. So does anyone have any thoughts on this, im pretty stuck and could use some input. https://www.dropbox.com/s/28fmx3oyyq...OA_mach5_2.zip

May 15, 2013, 07:18		#34
FJSJ Senior Member Javi Join Date: Jan 2013 Posts: 276 Rep Power: 16	I´ve just used the "associate edge to surface" tool and "associate face to surface" tool to resolve your problem in your outlet part. However, I´ve checked your quality determinant 3x3x3 and I think I´ve seen bad elements... https://www.dropbox.com/s/63go4kbobm...OA_mach5_2.rar

May 20, 2013, 04:58		#35
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	I missed this that last post somehow but thanks for that FJSJ. It seemed to work a treat. I played around a bit and got the quality up. I got pretty much all the elements above 0.9 for the 3x3x3 determinate, though there are are couple of sketchy elements at the trailing edge mostly (and a handful at the wing tip) Problem is now I am getting this weird error when I go to import my mesh into CFX. And it wont load up any of the domains or anything. I have tried renumbering the mesh and creating new parts in attempt to fix the issue but I havent been sucessful yet. Im so close to being able to run a sim that I can taste it, but have been beaten again! Does anyone have any ideas? latest files: https://www.dropbox.com/s/c972geny1f...oa_mach5_5.zip

May 20, 2013, 20:06		#36
ShowponyStuart Member Stuart Join Date: Jun 2012 Posts: 59 Rep Power: 13	No one has any thoughts on how to fix this error?