Compression stoke is giving higher pressure than calculated

nickjuana · April 20, 2015, 22:51

The only way is to do hexa mesh..? The geometry is some what complex

ghorrocks · April 20, 2015, 23:21

The mesh motion will run a lot better and more reliably if the bit which squashes up is a hex mesh (or a wedge mesh, anything which is an extrusion in the direction of squashing).

So in this case I would put the features at the bottom of your image in one mesh region which can be a tet mesh with inflation, but then transition to an extruded mesh for the cylindrical potion. Join the two with an interface with a 1:1 connection. Then make the tet mesh region either stationary or move with a velocity everywhere so this section just translates, and put the mesh motion on the cylindrical region so only this region squashes up.

I have done this many times when I used to do IC engine simulations in CFX so I can assure you it works well. I had features on both the piston and the cylinder head with a cylindrical region connecting them, and all the mesh squashing was done in the cylindrical section.

nickjuana · April 22, 2015, 08:28

Are you telling me that put outer surface as tetra and inner region as hexa..?

ghorrocks · April 22, 2015, 21:34

No. Here's a dodgy drawing of what I mean on your geometry:

image.png

The cylindrical bit is an extruded mesh (probably hex, but could be wedges). The complex bit is a tet/inflation layer mesh.

nickjuana · April 23, 2015, 00:18

Is it possible to do in ICEM CFD ..? or only with workbench
Thing is that I want hexa mesh internally for bottom portion that is the piston profile then only I will be able to compress.

ghorrocks · April 23, 2015, 05:55

This is easy to do in ICEM. ICEM is very flexible.

It is possible to do it in Workbench but not as easily.

You can mesh the piston profile any way you like. Just use an extruded mesh (hex or wedge) for the section which you squash and stretch to do the piston motion.

nickjuana · May 3, 2015, 12:54

I tried to revolve mesh by modelling only 60 degree section. I tried to define periodicity. I enabled mesh and tried to periodic, after periodic for a particular vertice mesh disappears. One particular periodic vertices is making mesh disappear

ghorrocks · May 3, 2015, 19:11

I do not understand you question. Please post an image of what you are seeing.

nickjuana · May 4, 2015, 00:21

This is geometry I am trying to revolve and I am trying to define periodicity with mesh enabled.

After clicking next vertices, the mesh is not generating. There are two such vertices creating problem. Which are preventing the mesh. what to do..?

ghorrocks · May 4, 2015, 00:33

This is an ICEM problem, not CFX. Try the geometry and meshing forum.

But I can see your blocking approach is not good. You want your mesh to match up across the periodic boundary so you should use a blocking which links across the boundaries. It is also much more complex than it needs to be from what I can see. Can you simplify it and use less blocks?

nickjuana · May 4, 2015, 02:05

I used Y grid at right side because it is complex geometry. It is some what complex profile so thats why I divided into many blocks.

ghorrocks · May 4, 2015, 06:44

Sure, but as you are using periodic boundaries this creates an additional constraint on the way you block it so the mesh matches on both sides of the interface. You will need to think about that as well as the blocking to fit into the geometry.

Alternately you could forget about matching the mesh over the periodic boundary and just connect it with a GGI.

nickjuana · May 5, 2015, 12:44

Can you please tell about GGI...?
Because I am unable to find solution to revolve this complex geometry

ghorrocks · May 5, 2015, 23:13

It is in the interface tab. You should have a choice of 1:1 or GGI.

nickjuana · May 6, 2015, 22:33

Geometry

Problem area

some mesh are coming inside and after solving few iteration its saying mesh is folded or invalid

ghorrocks · May 7, 2015, 02:25

You have not implemented the recommendation I told you in my post #44. That is the way to fix this.

You also need to be sure your mesh is correct with the blocks connecting properly (Ie contiguous)

nickjuana · May 7, 2015, 02:48

I tried my level best to block this complex geometry but what you told in post44, I dont know how to do that because piston with profile on the land part should have Hexa mesh. How do I put hexa interior and tetra exterior in the piston part...?

ghorrocks · May 7, 2015, 18:17

You are always going to have problems moving this mesh unless you implement my suggestion. So you need to do some meshing tutorials so you can see how to hex mesh these geometries and split it into multiple mesh blocks.

nickjuana · May 8, 2015, 00:23

Brother I am ready to do in your way but I have no idea how to do that. Is it like put only tetra in piston and hexa in cylinder....?
I tried putting tetra in piston alone but some mesh are coming out of the piston. what to do in that case ..?

ghorrocks · May 8, 2015, 03:10

Meshing this sort of object is easy in ICEM. You can easily do one section as a inflation/tet mesh and the other as a swept mesh. But it is a total pain in Workbench mesher. I am sure it is possible but it always causes me grief when I try to do something like this.

April 23, 2015, 00:18		#45
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	Is it possible to do in ICEM CFD ..? or only with workbench Thing is that I want hexa mesh internally for bottom portion that is the piston profile then only I will be able to compress. Last edited by nickjuana; April 23, 2015 at 02:59.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pulsatile pressure inlet with pressure outlet	a.lynchy	FLUENT	3	March 23, 2012 13:45
UDF to define or adjust pressure???	engahmed	FLUENT	0	July 6, 2010 17:19
Neumann pressure BC and velocity field	Antech	Main CFD Forum	0	April 25, 2006 02:15
Gas pressure question	Dan Moskal	Main CFD Forum	0	October 24, 2002 22:02
Terrible Mistake In Fluid Dynamics History	Abhi	Main CFD Forum	12	July 8, 2002 09:11

April 20, 2015, 22:51		#41
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	The only way is to do hexa mesh..? The geometry is some what complex

April 20, 2015, 23:21		#42
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	The mesh motion will run a lot better and more reliably if the bit which squashes up is a hex mesh (or a wedge mesh, anything which is an extrusion in the direction of squashing). So in this case I would put the features at the bottom of your image in one mesh region which can be a tet mesh with inflation, but then transition to an extruded mesh for the cylindrical potion. Join the two with an interface with a 1:1 connection. Then make the tet mesh region either stationary or move with a velocity everywhere so this section just translates, and put the mesh motion on the cylindrical region so only this region squashes up. I have done this many times when I used to do IC engine simulations in CFX so I can assure you it works well. I had features on both the piston and the cylinder head with a cylindrical region connecting them, and all the mesh squashing was done in the cylindrical section.

April 22, 2015, 08:28		#43
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	Are you telling me that put outer surface as tetra and inner region as hexa..?

April 22, 2015, 21:34		#44
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	No. Here's a dodgy drawing of what I mean on your geometry: image.png The cylindrical bit is an extruded mesh (probably hex, but could be wedges). The complex bit is a tet/inflation layer mesh.

April 23, 2015, 05:55		#46
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	This is easy to do in ICEM. ICEM is very flexible. It is possible to do it in Workbench but not as easily. You can mesh the piston profile any way you like. Just use an extruded mesh (hex or wedge) for the section which you squash and stretch to do the piston motion.

May 3, 2015, 12:54		#47
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	I tried to revolve mesh by modelling only 60 degree section. I tried to define periodicity. I enabled mesh and tried to periodic, after periodic for a particular vertice mesh disappears. One particular periodic vertices is making mesh disappear

May 3, 2015, 19:11		#48
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	I do not understand you question. Please post an image of what you are seeing.

May 4, 2015, 00:21		#49
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	This is geometry I am trying to revolve and I am trying to define periodicity with mesh enabled. After clicking next vertices, the mesh is not generating. There are two such vertices creating problem. Which are preventing the mesh. what to do..?

May 4, 2015, 00:33		#50
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	This is an ICEM problem, not CFX. Try the geometry and meshing forum. But I can see your blocking approach is not good. You want your mesh to match up across the periodic boundary so you should use a blocking which links across the boundaries. It is also much more complex than it needs to be from what I can see. Can you simplify it and use less blocks?

May 4, 2015, 02:05		#51
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	I used Y grid at right side because it is complex geometry. It is some what complex profile so thats why I divided into many blocks.

May 4, 2015, 06:44		#52
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Sure, but as you are using periodic boundaries this creates an additional constraint on the way you block it so the mesh matches on both sides of the interface. You will need to think about that as well as the blocking to fit into the geometry. Alternately you could forget about matching the mesh over the periodic boundary and just connect it with a GGI.

May 5, 2015, 12:44		#53
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	Can you please tell about GGI...? Because I am unable to find solution to revolve this complex geometry

May 5, 2015, 23:13		#54
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	It is in the interface tab. You should have a choice of 1:1 or GGI.

May 6, 2015, 22:33		#55
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	Geometry Problem area some mesh are coming inside and after solving few iteration its saying mesh is folded or invalid

May 7, 2015, 02:25		#56
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	You have not implemented the recommendation I told you in my post #44. That is the way to fix this. You also need to be sure your mesh is correct with the blocks connecting properly (Ie contiguous)

May 7, 2015, 02:48		#57
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	I tried my level best to block this complex geometry but what you told in post44, I dont know how to do that because piston with profile on the land part should have Hexa mesh. How do I put hexa interior and tetra exterior in the piston part...?

May 7, 2015, 18:17		#58
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	You are always going to have problems moving this mesh unless you implement my suggestion. So you need to do some meshing tutorials so you can see how to hex mesh these geometries and split it into multiple mesh blocks.

May 8, 2015, 00:23		#59
nickjuana Member Nick Join Date: Apr 2015 Posts: 40 Rep Power: 11	Brother I am ready to do in your way but I have no idea how to do that. Is it like put only tetra in piston and hexa in cylinder....? I tried putting tetra in piston alone but some mesh are coming out of the piston. what to do in that case ..?

May 8, 2015, 03:10		#60
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Meshing this sort of object is easy in ICEM. You can easily do one section as a inflation/tet mesh and the other as a swept mesh. But it is a total pain in Workbench mesher. I am sure it is possible but it always causes me grief when I try to do something like this.