Meshfree CFD

meandcfd · July 10, 2019, 10:15

I agree that automatic mesh helps reduce time but I'm doing a simulation of a rotating high-powered spray of liquid inside of a pipe. It took weeks of work using CFX and did not come close to physical tests. The particle method was much faster and closer to the physical. Running a more detailed study now to see how close we can get in simulation.

flotus1 · July 10, 2019, 13:43

Out of curiosity, and I should probably have asked this earlier: which part of meshing took weeks? Manually generating a hexa mesh for the given geometry? Or did an auto-mesher take this long? Which software was used?

meandcfd · July 10, 2019, 15:22

Ansys Meshing
The problem is that I still have to fix / adjust lots of areas in the mesh. And then the solve time is really long. Particle method was faster for setup and solve but we need to verify accuracy. Mesh or particles depends on the problem you need to address doesn't it?

flotus1 · July 11, 2019, 01:19

IDK, literal weeks for auto-generating a tet/poly mesh with the help of an auto mesher sounds excessive. Turnover times should rather be less than a day for an experienced user, maybe a bit more for very complex geometries.
I never used Ansys meshers for complex flow problems, but there sure are a lot of dials and switches to change its behavior.
Of course I can't comment on your particular flow case, but changing the numerical method because grid generation took too long sounds like you skipped a few steps. Like properly preparing geometry and surfaces for the auto mesher, getting it dialed in for your particular problem (maybe with the help of support), or switching to a different mesher.
I am not saying the approach IS wrong in your particular case, just that based on the information presented here I would have investigated the grid generation stage first.

Quote:

Mesh or particles depends on the problem you need to address doesn't it?

Exactly. Which method you use should depend more on the type of flow you have at hand. And less on how easy it is to get a decent mesh.

meandcfd · July 11, 2019, 10:21

I wasn't clear - It didn't take weeks generating a mesh. The project took weeks of work using CFX, some of which was spent on prepping the model so it could be meshed.

But you make my point - "properly preparing geometry and surfaces for the automesh" - why? because the mesh generators require it. But if you can that take existing geometry, apply a different solver and get answers that match physical test, then why mesh?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFD Online Celebrates 20 Years Online	jola	Site News & Announcements	22	January 31, 2015 00:30
STAR-Works : Mainstream CAD with CFD	CD adapco Group Marketing	Siemens	0	February 13, 2002 12:23
Where do we go from here? CFD in 2001	John C. Chien	Main CFD Forum	36	January 24, 2001 21:10
ASME CFD Symposium, Atlanta, July 2001	Chris R. Kleijn	Main CFD Forum	0	August 21, 2000 04:49
Which is better to develop in-house CFD code or to buy a available CFD package.	Tareq Al-shaalan	Main CFD Forum	10	June 12, 1999 23:27

July 10, 2019, 10:15		#21
meandcfd New Member Join Date: Jun 2018 Posts: 21 Rep Power: 7	I agree that automatic mesh helps reduce time but I'm doing a simulation of a rotating high-powered spray of liquid inside of a pipe. It took weeks of work using CFX and did not come close to physical tests. The particle method was much faster and closer to the physical. Running a more detailed study now to see how close we can get in simulation.

July 10, 2019, 13:43		#22
flotus1 Super Moderator Alex Join Date: Jun 2012 Location: Germany Posts: 3,399 Rep Power: 46	Out of curiosity, and I should probably have asked this earlier: which part of meshing took weeks? Manually generating a hexa mesh for the given geometry? Or did an auto-mesher take this long? Which software was used?

July 10, 2019, 15:22		#23
meandcfd New Member Join Date: Jun 2018 Posts: 21 Rep Power: 7	Ansys Meshing The problem is that I still have to fix / adjust lots of areas in the mesh. And then the solve time is really long. Particle method was faster for setup and solve but we need to verify accuracy. Mesh or particles depends on the problem you need to address doesn't it?

July 11, 2019, 10:21		#25
meandcfd New Member Join Date: Jun 2018 Posts: 21 Rep Power: 7	I wasn't clear - It didn't take weeks generating a mesh. The project took weeks of work using CFX, some of which was spent on prepping the model so it could be meshed. But you make my point - "properly preparing geometry and surfaces for the automesh" - why? because the mesh generators require it. But if you can that take existing geometry, apply a different solver and get answers that match physical test, then why mesh?