Pressure drop too less

abcdefgh · March 16, 2015, 08:01

Now I've tried to do the SAS-SST Simulation.
But I think my mesh is not good enough.
Do you have an Idea, how i can make it better with that complex geometry?
The Problem is, i can't create a Prism Layer with a good Quality. A also the Mesh Expension Faktor is always pretty bad (120!).
If I make it like in the Icem Tutorials (First Octree with Prism, then smooth with Penta freeze, then Delaunay) I don't get good quality. What would you recommend?
Thank you!

Thomas MADELEINE · March 17, 2015, 05:03

Hi all,
just one thing I have noticed in the first post. I have read the others but quite quickly so excuse me if it was already mentioned.

Why do you model your fluid (air) as an incompressible gas ? Does it stay in quite low velocities ?

abcdefgh · March 17, 2015, 05:08

The velocities are very low. 2-5 m/s. So the Ma is also very low.

ghorrocks · March 17, 2015, 05:18

A common problem with things like touching spheres is that they meeting at a tangent point. To mesh this tangent point the mesher needs to put horribly distorted elements in to mesh the fine cusp near the tangent point.

A few simple work arounds are:
* Make the spheres penetrate each other slightly.
* Put a small fillet radii around the tangent point.
* Separate the spheres by a small amount.

All of these methods will make a big difference in the mesh quality, with the cost of a small modification to the geometry.

abcdefgh · March 17, 2015, 05:55

Thank you!
But my Geometry is already prepared: I put inside between two spheres litte cylinders from the Middlepoints of the spheres. So the spheres are all connected with "bridges" to all theire neighbour sphere. Thats a Geometry at which it should be possible to make a good mesh.

ghorrocks · March 17, 2015, 06:29

I do not completely understand what you are saying - I think you are saying you have already implemented what I suggested.

If that is the case, then make the cylinders larger! That will improve mesh quality more.

abcdefgh · March 27, 2015, 11:05

Hey.

to make the cylinders larger would produce such a big change oft the Geometry. The cylinders are at the moment 2 mm, the spheres are 10 mm. Don't you think that it is enough? But its really a Problem for me to make a good mesh. The mesh expension Factor is always very high (>50-100) and Prism I can't do inside. Do you have an Idea how I can make it better? Tutorials can't help

I didn't find the mistake yet for the Problem with the pressure. I don't think that it is only the mesh.
My velocities are not high (Ma<0,15) but I want to try it with simulating a compressible flow, what somebody of you asked me.
I used air ideal gas (can I use it with higher temperatures or only with 25 °C???) and in the Fluid model I changed from thermal to total energy.
But the result is the same.
The flow of the Inlet and Oulet is subsonic. But there was an instability during the calculation. I also tried it with mixed/supersoncic...The result was nearly the same but the it was stable during the calculation.

The next wrong thing with an incompressible flow: in my simulation the pressure drop becomes lower with higher temperatures. That's not right: It should become higher!!!!

ghorrocks · March 28, 2015, 04:56

Compressible flow will not help. If the flow is incompressible then model it using an incompressible model.

Your problem is mesh quality (and possibly mesh resolution). I do not think you understand my suggestion. Have a look at the poor quality elements in your simulation. You will find they all occur near the tangential touching points of the spheres on each other. The mesher is trying to squeeze very distorted elements into this sharp cusp.

So you should remove the cusp. You can do this by making the spheres overlap a small amount, or putting a fillet around the tangent points. This is a modification of the geometry but only of a few microns and only in very wall constrained flow - so that means the modification will not affect results significantly if done properly.

Can you explain how the pressure drop becomes higher with higher temperature? What causes this? At higher temperature the viscosity will become lower (leading to less pressure drop) and the density will become lower (also leading to less pressure drop). So something else must be causing the pressure drop to increase with increasing temperature.

abcdefgh · March 30, 2015, 01:39

You are right. Thats not the solution.
I had prepared the Geometry with cylinders between the middle points of the spheres (picture). There the Geometry change is the smallest.
Isn't that enough?
The way of my meshing:
Octree Volume mesh
Deleting volume mesh
Smooth the surface mesh
Delauney volume

Smooth that volume mesh
Prism layers with an expansion ratio of 1.2, but no initial height set

Smooth tetras freezing pentas and pyramids
smooth everything

The viscoity of air becomes higher with higher temperatures I think!
And thats also what my measurements say.

ghorrocks · March 30, 2015, 03:15

Sorry, you are correct. Air viscosity increases with temperature.

If you cannot get the mesh quality adequate with the cylinders you have then you will have to make them bigger. This is going to be a tricky thing to get a good mesh on.

abcdefgh · March 30, 2015, 05:43

At the moment the cylinders between the spheres are 2mm, the spheres are 10 mm (picture). I made a small radius on it, that the transition is not so hard. Would it be better without? What do you think?
And is the Way how ich mesh it the right one?
Octree Volume mesh
Deleting volume mesh
Smooth the surface mesh
Delauney volume
Smooth that volume mesh
Prism layers with an expansion ratio of 1.2 on wall, spheres and cylinders, but no initial height set
Smooth tetras freezing pentas and pyramids
smooth everything

Then checking and look at the Quality (Skewness, Aspect ratio, Orthongonality).

ghorrocks · March 30, 2015, 05:48

Whether the fillet is helping will depend on the size mesh you are putting in it. If you are putting a relatively coarse mesh in it then no fillets is better. If you are putting a fine mesh in it then it will be better with the fillets. By "fine" I mean whether the fillets are being resolved by the mesh or not.

Be careful with mesh smoothing in geometries like this. It can do more harm than good. Try remeshing with no mesh smoothing.

But you are just going to have to try some options and see what works. Meshing is an art which will only improve with practise.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Pressure drop using Fan type BC	Alexis Sack	OpenFOAM Running, Solving & CFD	2	September 22, 2014 10:18
How to study pressure drop of continous phase in VOF model	sajeesh	FLUENT	4	February 5, 2014 23:01
pressure drop in the pipe	mike.mm	FLUENT	6	February 7, 2011 11:31
Pipe Flow - Pressure Drop	Daniel L	FLOW-3D	2	December 10, 2010 05:23
Pressure Drop at entrance of a rotor-stator.	Resnick	Main CFD Forum	0	November 20, 2007 15:50

March 16, 2015, 08:01		#21
abcdefgh New Member Join Date: Oct 2014 Posts: 24 Rep Power: 12	Now I've tried to do the SAS-SST Simulation. But I think my mesh is not good enough. Do you have an Idea, how i can make it better with that complex geometry? The Problem is, i can't create a Prism Layer with a good Quality. A also the Mesh Expension Faktor is always pretty bad (120!). If I make it like in the Icem Tutorials (First Octree with Prism, then smooth with Penta freeze, then Delaunay) I don't get good quality. What would you recommend? Thank you!

March 17, 2015, 05:03		#22
Thomas MADELEINE Senior Member Thomas MADELEINE Join Date: Oct 2014 Posts: 126 Rep Power: 12	Hi all, just one thing I have noticed in the first post. I have read the others but quite quickly so excuse me if it was already mentioned. Why do you model your fluid (air) as an incompressible gas ? Does it stay in quite low velocities ?

March 17, 2015, 05:08		#23
abcdefgh New Member Join Date: Oct 2014 Posts: 24 Rep Power: 12	The velocities are very low. 2-5 m/s. So the Ma is also very low.

March 17, 2015, 05:18		#24
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	A common problem with things like touching spheres is that they meeting at a tangent point. To mesh this tangent point the mesher needs to put horribly distorted elements in to mesh the fine cusp near the tangent point. A few simple work arounds are: * Make the spheres penetrate each other slightly. * Put a small fillet radii around the tangent point. * Separate the spheres by a small amount. All of these methods will make a big difference in the mesh quality, with the cost of a small modification to the geometry.

March 17, 2015, 05:55		#25
abcdefgh New Member Join Date: Oct 2014 Posts: 24 Rep Power: 12	Thank you! But my Geometry is already prepared: I put inside between two spheres litte cylinders from the Middlepoints of the spheres. So the spheres are all connected with "bridges" to all theire neighbour sphere. Thats a Geometry at which it should be possible to make a good mesh.

March 17, 2015, 06:29		#26
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	I do not completely understand what you are saying - I think you are saying you have already implemented what I suggested. If that is the case, then make the cylinders larger! That will improve mesh quality more.

March 27, 2015, 11:05		#27
abcdefgh New Member Join Date: Oct 2014 Posts: 24 Rep Power: 12	Hey. to make the cylinders larger would produce such a big change oft the Geometry. The cylinders are at the moment 2 mm, the spheres are 10 mm. Don't you think that it is enough? But its really a Problem for me to make a good mesh. The mesh expension Factor is always very high (>50-100) and Prism I can't do inside. Do you have an Idea how I can make it better? Tutorials can't help I didn't find the mistake yet for the Problem with the pressure. I don't think that it is only the mesh. My velocities are not high (Ma<0,15) but I want to try it with simulating a compressible flow, what somebody of you asked me. I used air ideal gas (can I use it with higher temperatures or only with 25 °C???) and in the Fluid model I changed from thermal to total energy. But the result is the same. The flow of the Inlet and Oulet is subsonic. But there was an instability during the calculation. I also tried it with mixed/supersoncic...The result was nearly the same but the it was stable during the calculation. The next wrong thing with an incompressible flow: in my simulation the pressure drop becomes lower with higher temperatures. That's not right: It should become higher!!!!

March 28, 2015, 04:56		#28
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	Compressible flow will not help. If the flow is incompressible then model it using an incompressible model. Your problem is mesh quality (and possibly mesh resolution). I do not think you understand my suggestion. Have a look at the poor quality elements in your simulation. You will find they all occur near the tangential touching points of the spheres on each other. The mesher is trying to squeeze very distorted elements into this sharp cusp. So you should remove the cusp. You can do this by making the spheres overlap a small amount, or putting a fillet around the tangent points. This is a modification of the geometry but only of a few microns and only in very wall constrained flow - so that means the modification will not affect results significantly if done properly. Can you explain how the pressure drop becomes higher with higher temperature? What causes this? At higher temperature the viscosity will become lower (leading to less pressure drop) and the density will become lower (also leading to less pressure drop). So something else must be causing the pressure drop to increase with increasing temperature.

March 30, 2015, 01:39		#29
abcdefgh New Member Join Date: Oct 2014 Posts: 24 Rep Power: 12	You are right. Thats not the solution. I had prepared the Geometry with cylinders between the middle points of the spheres (picture). There the Geometry change is the smallest. Isn't that enough? The way of my meshing: Octree Volume mesh Deleting volume mesh Smooth the surface mesh Delauney volume Smooth that volume mesh Prism layers with an expansion ratio of 1.2, but no initial height set Smooth tetras freezing pentas and pyramids smooth everything The viscoity of air becomes higher with higher temperatures I think! And thats also what my measurements say.

March 30, 2015, 03:15		#30
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	Sorry, you are correct. Air viscosity increases with temperature. If you cannot get the mesh quality adequate with the cylinders you have then you will have to make them bigger. This is going to be a tricky thing to get a good mesh on.

March 30, 2015, 05:48		#32
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,852 Rep Power: 144	Whether the fillet is helping will depend on the size mesh you are putting in it. If you are putting a relatively coarse mesh in it then no fillets is better. If you are putting a fine mesh in it then it will be better with the fillets. By "fine" I mean whether the fillets are being resolved by the mesh or not. Be careful with mesh smoothing in geometries like this. It can do more harm than good. Try remeshing with no mesh smoothing. But you are just going to have to try some options and see what works. Meshing is an art which will only improve with practise.