Tetra Meshing over Frigate Ship Model

rohitsaini@y7mail.com · August 29, 2012, 14:06

I want to analyse the vortex shedding from a Frigate ship model for which I need to apply tetra-prism meshing. I am new to ICEM, can anyone please help me out in meshing such that i get fine meshing around the ship model and coarse in the surrounding volume, what would be the initial mesh parameters and how should I proceed to get the good result. (Dimension of control volume is length = 600 ft, width = 175 ft, height = 150 ft)

Felggv · August 29, 2012, 15:42

Hi there,

As far as I know, it's better for you to mesh them separatedly.

First mesh your ship the way you want, then mesh the surrounding volume.

After you've done that, you simply create an interface between them when editing your BC's in CFX-Pre. That's a way to link them!

Good luck.

ghorrocks · August 29, 2012, 19:35

How do you mesh a complex geometry? As if we can answer that on a forum.

Do the tutorials (including the tutorials at the ANSYS Customer page), do training, play with lots of options and see how they go, and do benchmark studies to see what the accuracy implications for meshing choices are. There is not short cuts, you just have to do the work to become proficient.

evcelica · August 29, 2012, 21:42

your geometry looks to be blocks, and all 90 degree corners. I'm no expert but that seems like a perfect candidate for a perfectly orthogonal hexa mesh?

rohitsaini@y7mail.com · August 30, 2012, 07:33

I have done terta meshing with cell density around the ship and exponential meshing parameter around the ship with coarse meshing. When I am analysing the steady solution in the Fluent, My iteration graph is oscillating even in the k-epsioln realisable model at 1st order mode. Please guys help me out with the reason behind the oscillation if the graph.

singer1812 · August 30, 2012, 10:14

Umm, this is a cfx forum. Perhaps fluent forum could address better?

Your domain looks very blocky. Are you sure you dont want to use ICEM blocking for this?

If you insist on using tet, perhaps you are expanding too fast from the fine region to the coarse. Perhaps you still have too coarse around the frigate. Did you do a mesh sensitivity to try to get the convergence you want?

Felggv · August 30, 2012, 14:05

This is a simple strategy (somehow).

You don't know the problem, but you may know what is NOT causing the problem... so go around every parameter and try to eliminate them by steps, you'll get closer the what's causing you trouble by attacking isolating the influence of the other parameters.

This works for most cases.

Good luck!

PS: I'd say the first thing you should do is the mesh analysis.

rohitsaini@y7mail.com · August 31, 2012, 08:11

I have done hexa meshing with blocking technique and found my mesh quality is 0.7. Is it quite enough for this model and Please tell me the way, how should I give ratio and spacing such that my aspect ratio can be maintained.

Felggv · August 31, 2012, 10:20

Hi,

Let me ask you, have you set the dimensions of your model in a way that it would fit the element size you want? Do the blocks fit?

Let's say you want a 4cm element size mesh... then you need your dimensions to be proportional to 4!

Keeping that in mind might not give you the real dimensions of your frigate but would be a nice initial step for solving the problem... I've seen this method giving mesh quality of ~1.0 and, besides, mesh would be out of question... one thing not to worry about anymore.

Good luck!

rohitsaini@y7mail.com · August 31, 2012, 10:41

hey Felggv, thanx for your precious help...

rohitsaini@y7mail.com · August 31, 2012, 14:11

I split the blocks and done all the required association and meshing part, but when I am analyzing the mesh in Fluent, then the graph is oscillating even upto 10000 iterations, I think, this is the problem with meshing, aspect ratio might be not good so while examining meshing through interpolation it is oscillating. I still think there is some problem with ratio's and spacing. Is I am rite?

Felggv · September 3, 2012, 15:26

Try analyzing the mesh within the ICEM.

It has some quality tools there... the goal is to achieve something close to 1.
Keep trying different ratios to see the overall quality, although your model looks quite simple to mesh (using blocks), it can be a little tricky because of the distortion on the inside caused by any wrong spacing.

Are you doing separated meshes for the frigate itself and the atmosphere around it?

Cheers.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[GAMBIT] pro\e model to be exported for gambit meshing	mazhar1613	ANSYS Meshing & Geometry	1	August 18, 2013 14:36
icem cfd tetra meshing error	joegi.geo	ANSYS Meshing & Geometry	6	July 29, 2011 12:00
Tetra Meshing. 2d or 3d?	syler3321	Main CFD Forum	0	June 29, 2010 14:04
Octree Vs Advancing Front tetra meshing techniques	Sam	CFX	4	February 18, 2008 09:54
DPM model w/ Wave model - errors in documentation	HS	FLUENT	0	April 12, 2006 05:37

August 29, 2012, 15:42		#2
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Hi there, As far as I know, it's better for you to mesh them separatedly. First mesh your ship the way you want, then mesh the surrounding volume. After you've done that, you simply create an interface between them when editing your BC's in CFX-Pre. That's a way to link them! Good luck.

August 29, 2012, 19:35		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	How do you mesh a complex geometry? As if we can answer that on a forum. Do the tutorials (including the tutorials at the ANSYS Customer page), do training, play with lots of options and see how they go, and do benchmark studies to see what the accuracy implications for meshing choices are. There is not short cuts, you just have to do the work to become proficient.

August 29, 2012, 21:42		#4
evcelica Senior Member Erik Join Date: Feb 2011 Location: Earth (Land portion) Posts: 1,188 Rep Power: 23	your geometry looks to be blocks, and all 90 degree corners. I'm no expert but that seems like a perfect candidate for a perfectly orthogonal hexa mesh?

August 30, 2012, 07:33		#5
rohitsaini@y7mail.com New Member Rohit Saini Join Date: Jul 2012 Posts: 13 Rep Power: 14	I have done terta meshing with cell density around the ship and exponential meshing parameter around the ship with coarse meshing. When I am analysing the steady solution in the Fluent, My iteration graph is oscillating even in the k-epsioln realisable model at 1st order mode. Please guys help me out with the reason behind the oscillation if the graph.

August 30, 2012, 10:14		#6
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 21	Umm, this is a cfx forum. Perhaps fluent forum could address better? Your domain looks very blocky. Are you sure you dont want to use ICEM blocking for this? If you insist on using tet, perhaps you are expanding too fast from the fine region to the coarse. Perhaps you still have too coarse around the frigate. Did you do a mesh sensitivity to try to get the convergence you want?

August 30, 2012, 14:05		#7
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	This is a simple strategy (somehow). You don't know the problem, but you may know what is NOT causing the problem... so go around every parameter and try to eliminate them by steps, you'll get closer the what's causing you trouble by attacking isolating the influence of the other parameters. This works for most cases. Good luck! PS: I'd say the first thing you should do is the mesh analysis.

August 31, 2012, 08:11		#8
rohitsaini@y7mail.com New Member Rohit Saini Join Date: Jul 2012 Posts: 13 Rep Power: 14	I have done hexa meshing with blocking technique and found my mesh quality is 0.7. Is it quite enough for this model and Please tell me the way, how should I give ratio and spacing such that my aspect ratio can be maintained.

August 31, 2012, 10:20		#9
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Hi, Let me ask you, have you set the dimensions of your model in a way that it would fit the element size you want? Do the blocks fit? Let's say you want a 4cm element size mesh... then you need your dimensions to be proportional to 4! Keeping that in mind might not give you the real dimensions of your frigate but would be a nice initial step for solving the problem... I've seen this method giving mesh quality of ~1.0 and, besides, mesh would be out of question... one thing not to worry about anymore. Good luck!

August 31, 2012, 10:41		#10
rohitsaini@y7mail.com New Member Rohit Saini Join Date: Jul 2012 Posts: 13 Rep Power: 14	hey Felggv, thanx for your precious help...

August 31, 2012, 14:11		#11
rohitsaini@y7mail.com New Member Rohit Saini Join Date: Jul 2012 Posts: 13 Rep Power: 14	I split the blocks and done all the required association and meshing part, but when I am analyzing the mesh in Fluent, then the graph is oscillating even upto 10000 iterations, I think, this is the problem with meshing, aspect ratio might be not good so while examining meshing through interpolation it is oscillating. I still think there is some problem with ratio's and spacing. Is I am rite?

September 3, 2012, 15:26		#12
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Try analyzing the mesh within the ICEM. It has some quality tools there... the goal is to achieve something close to 1. Keep trying different ratios to see the overall quality, although your model looks quite simple to mesh (using blocks), it can be a little tricky because of the distortion on the inside caused by any wrong spacing. Are you doing separated meshes for the frigate itself and the atmosphere around it? Cheers.