[ICEM] Surfaces intersecting after removing o-grid block

rich85 · April 23, 2011, 09:37

Hey

Not sure how accurate the title is but its best I could think of. From the attached pictures are the solid model, the models curves in ICEM along with the model blocked and meshed and what the mesh looks like in CFX Pre.

I have been following the 3D pipe junction tutorial as the geometry is similar and my problems start when I start removing blocks to give me the hollow sections. The model I've made is of the fluid, in this case air of a gamma type Stirling engine, the hollowed section is the displacer piston which will move in the analysis.

I'm looking at splitting the model up into different sections and mesh them individually but I am not sure whether to then merge he meshes within ICEM or within CFX Pre.

I'm using ICEM so I could take advantage of the replay feature to re-mesh the model when distortion reaches a set value as I have tried CFX Mesh in workbench to create the mesh but after a couple of time steps I get a negative volume error due to the predominance of tet elements.

If anyone has any tips, pointers or alternatives on how to get a suitable mesh to perform a analysis with the pistons moving it would be greatly appreciated.

Rich

PSYMN · April 25, 2011, 16:10

Your last image seems to show the blocking, but it is not a good topology at all...

Just looking at the cylinder on the right... What you really have are three cylinders...

1) The outer wall

2) The chamber wall

3) the valve

Just imagine these extended from end to end. That is how you should block it. One Ogrid (CGrid since it is a symmetry model) that has its central HGrid within the valve stem. Then split the Ogrid twice (Once for the valve diameter and once for the piston diameter). Split along the cylinder for the piston face... Make all your associations and delete the blocks that are not part of your fluid volume...

rich85 · April 26, 2011, 07:46

Thanks for the help, i've got the mesh working now.

However when starting solver it exits with return code 1 and just before i get a notice about the heat transfer coefficient uses wall adjacent temperature for bulk temperature but no error code or other message indicating whats wrong.

I've changed a few of the boundary conditons within pre but i'm now out of ideas.

PSYMN · April 26, 2011, 10:05

Sorry, not sure about your solver issue. Perhaps ask the question again on the Fluent forum? Be sure to include more of the actual error message.

April 23, 2011, 09:37	Surfaces intersecting after removing o-grid block	#1
rich85 New Member Richard Evans Join Date: Feb 2010 Posts: 2 Rep Power: 0	Hey Not sure how accurate the title is but its best I could think of. From the attached pictures are the solid model, the models curves in ICEM along with the model blocked and meshed and what the mesh looks like in CFX Pre. I have been following the 3D pipe junction tutorial as the geometry is similar and my problems start when I start removing blocks to give me the hollow sections. The model I've made is of the fluid, in this case air of a gamma type Stirling engine, the hollowed section is the displacer piston which will move in the analysis. I'm looking at splitting the model up into different sections and mesh them individually but I am not sure whether to then merge he meshes within ICEM or within CFX Pre. I'm using ICEM so I could take advantage of the replay feature to re-mesh the model when distortion reaches a set value as I have tried CFX Mesh in workbench to create the mesh but after a couple of time steps I get a negative volume error due to the predominance of tet elements. If anyone has any tips, pointers or alternatives on how to get a suitable mesh to perform a analysis with the pistons moving it would be greatly appreciated. Rich

April 25, 2011, 16:10		#2
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	Your last image seems to show the blocking, but it is not a good topology at all... Just looking at the cylinder on the right... What you really have are three cylinders... 1) The outer wall 2) The chamber wall 3) the valve Just imagine these extended from end to end. That is how you should block it. One Ogrid (CGrid since it is a symmetry model) that has its central HGrid within the valve stem. Then split the Ogrid twice (Once for the valve diameter and once for the piston diameter). Split along the cylinder for the piston face... Make all your associations and delete the blocks that are not part of your fluid volume... __________________ ----------------------------------------- Please help guide development at ANSYS by filling in these surveys Public ANSYS ICEM CFD Users Survey This second one is more general (Gambit, TGrid and ANSYS Meshing users welcome)... CFD Online Users Survey

April 26, 2011, 10:05		#4
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	Sorry, not sure about your solver issue. Perhaps ask the question again on the Fluent forum? Be sure to include more of the actual error message. __________________ ----------------------------------------- Please help guide development at ANSYS by filling in these surveys Public ANSYS ICEM CFD Users Survey This second one is more general (Gambit, TGrid and ANSYS Meshing users welcome)... CFD Online Users Survey

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Grid Independent Solution	Chuck Leakeas	Main CFD Forum	2	May 26, 2000 11:18
Algebraic Multigrid for nonuniform grid	Joshua Lee	Main CFD Forum	4	March 16, 2000 13:05
Non-uniform grid calculation	Aspens	Main CFD Forum	1	February 23, 2000 14:15
Grid Quality and the Solution	Faraz	Main CFD Forum	4	January 10, 2000 18:18
Numerical methods for discontinuous grid interfaces?	Hansong Hang	Main CFD Forum	12	September 16, 1998 22:26

April 26, 2011, 07:46		#3
rich85 New Member Richard Evans Join Date: Feb 2010 Posts: 2 Rep Power: 0	Thanks for the help, i've got the mesh working now. However when starting solver it exits with return code 1 and just before i get a notice about the heat transfer coefficient uses wall adjacent temperature for bulk temperature but no error code or other message indicating whats wrong. I've changed a few of the boundary conditons within pre but i'm now out of ideas.