CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] Volume Orientations Error In ICEM (http://www.cfd-online.com/Forums/ansys-meshing/120896-volume-orientations-error-icem.html)

ahmadreza July 17, 2013 01:47

Volume Orientations Error In ICEM
 
Hi
For My Master Project I need to have full structure Mesh on Ship Propeller


I can produce it but i have bad element (bad mesh) on tip of blade because of Leading edge of the blade is Sharp (Line)

Front of Leading edge I have Triangle Block ( I Know That is bad but i don't know How to modify it)

check Mesh ICEM 14.0 Shows:

ERROR:

the element is too bad and shows Volume Orientation Error

Please Help me How to Modify my Block

Thanks


http://upload7.ir/images/22350517418668597106.jpg
http://upload7.ir/images/08465106212636021821.jpg
http://upload7.ir/images/22447560473296158371.jpg
http://upload7.ir/images/35847535113813504727.jpg
http://upload7.ir/images/86102233848860599481.jpg
http://upload7.ir/images/54772143397342289193.png

FJSJ July 17, 2013 02:31

Hi,

Did you delete the blocks inside the propeller?

Is your trailing edge sharp or blunt? I think is the second one, but Im not sure.

I think you can improve the blocking. Take a look in this thread:

Ralen Approach

ahmadreza July 17, 2013 05:03

Thanks.

My Leading edge is sharp and my problem is here ///////////// and My trailing edge is blunt and it is good for me

(I write mistake tip propeller (MY tip Propeller is A little Blunk) )


Yes I Separate and delete Inside of Propeller


What your Idea about Block that is in the front of Picture

Regard
AhmadReza

PSYMN July 17, 2013 10:14

1 Attachment(s)
Instead of a triangle block against your sharp edge, what if you collapsed it? (collapse the other side of the triangle also, so it is a sharp block...

With ICEM CFD Hexa, the best advice is to think schematically.

In profile, your blade is rounded, so you should be thinking of a CGrid...

Basically, your model is a wedge that is a bit twisty... Is the tip sharp or blunt?

Your model is tricky because the transition from sharp to round probably happens gradually as the blocking goes around the corner. You may need to make a judgement call to approximate it somewhere and give up some accuracy in exchange for quality. For instance, if the transition happens along the tip, you may need to pretend that the tip is like a thin wedge with a sharp front, flat back and flat tip... You can either modify the geometry or turn off the projection.

Also, part of your problem is because you have not tidied up your blocking further out from the blade... You are forcing blocks to fit to a curvy blade on one side without adjusting the very square layout away from the blades. You can massage that into place a lot better... If the blocking were to twist around the blade a bit, it would improve the quality.

ahmadreza July 17, 2013 10:53

Hi
Thanks for your Answer

My Tip is a little blunt
My leading edge is Sharp
My Trailing edge have fillet
I cant collapse it because I have a rotary domain ( I need Two Domain For Define In Fluent & ) and I Produce A total block and I segment it to 4 block
and next time I segment whole of block to O-block
So I have 1 Block for 7 blades I segment it and


http://upload7.ir/images/21111571856490580570.png
http://upload7.ir/images/03975327150012423368.png
http://upload7.ir/images/42967238947214923277.png
http://upload7.ir/images/66186523225767739262.png

ahmadreza July 17, 2013 11:29

Quote:

Originally Posted by PSYMN (Post 440331)
Instead of a triangle block against your sharp edge, what if you collapsed it? (collapse the other side of the triangle also, so it is a sharp block...

With ICEM CFD Hexa, the best advice is to think schematically.

In profile, your blade is rounded, so you should be thinking of a CGrid...

Basically, your model is a wedge that is a bit twisty... Is the tip sharp or blunt?

Your model is tricky because the transition from sharp to round probably happens gradually as the blocking goes around the corner. You may need to make a judgement call to approximate it somewhere and give up some accuracy in exchange for quality. For instance, if the transition happens along the tip, you may need to pretend that the tip is like a thin wedge with a sharp front, flat back and flat tip... You can either modify the geometry or turn off the projection.

Also, part of your problem is because you have not tidied up your blocking further out from the blade... You are forcing blocks to fit to a curvy blade on one side without adjusting the very square layout away from the blades. You can massage that into place a lot better... If the blocking were to twist around the blade a bit, it would improve the quality.



There is not possible because of my blade have minimum 18 degree skew and rake and ... this mean is my blade can't model with a Collapse block

PSYMN July 17, 2013 12:24

I don't think that would stop you from collapsing the block, perhaps you do not understand what I mean... (Try the pipe blade tutorial for an example)

Also, I would suggest that you only model one segment of the problem (much less work) and then copy rotate it around to get the full geometry...

ahmadreza July 17, 2013 12:53

Quote:

Originally Posted by PSYMN (Post 440331)
Instead of a triangle block against your sharp edge, what if you collapsed it? (collapse the other side of the triangle also, so it is a sharp block...

With ICEM CFD Hexa, the best advice is to think schematically.

In profile, your blade is rounded, so you should be thinking of a CGrid...

Basically, your model is a wedge that is a bit twisty... Is the tip sharp or blunt?

Your model is tricky because the transition from sharp to round probably happens gradually as the blocking goes around the corner. You may need to make a judgement call to approximate it somewhere and give up some accuracy in exchange for quality. For instance, if the transition happens along the tip, you may need to pretend that the tip is like a thin wedge with a sharp front, flat back and flat tip... You can either modify the geometry or turn off the projection.

Also, part of your problem is because you have not tidied up your blocking further out from the blade... You are forcing blocks to fit to a curvy blade on one side without adjusting the very square layout away from the blades. You can massage that into place a lot better... If the blocking were to twist around the blade a bit, it would improve the quality.

Quote:

Originally Posted by PSYMN (Post 440359)
I don't think that would stop you from collapsing the block, perhaps you do not understand what I mean... (Try the pipe blade tutorial for an example)

Also, I would suggest that you only model one segment of the problem (much less work) and then copy rotate it around to get the full geometry...



thanks for Attention

My geometry is Contra Rotating propeller ( two propeller that rotate conter . the aft propeller are 5 blade and for propeller have 7 propeller ( the blade of for propeller are not same with aft propeller)

NOW
1-how i can make one blade or two blade and rotate ?

2- how i can make the farfeild forward and backward and Around of My Propeller?

3-How I can in this way make A cylinder around the propeller for produce two domain for Fluent?

PSYMN July 17, 2013 16:13

You will need 3 volumes...

The large outer volume will be stationary.

The two inner volumes can be discs that will be meshed separately (non conformal)... Instead of a full disc, you can mesh a single blade for each. Use periodicity and then copy rotate them around.

Each of the three disks will be a separate volume material (FLUID1, FLUID2 and FLUID3 or whatever sensible names you choose). In Fluent, you would set up each zone and the wall boundaries to get the desired rotation.

ahmadreza July 17, 2013 16:43

Quote:

Originally Posted by PSYMN (Post 440391)
You will need 3 volumes...

The large outer volume will be stationary.

The two inner volumes can be discs that will be meshed separately (non conformal)... Instead of a full disc, you can mesh a single blade for each. Use periodicity and then copy rotate them around.

Each of the three disks will be a separate volume material (FLUID1, FLUID2 and FLUID3 or whatever sensible names you choose). In Fluent, you would set up each zone and the wall boundaries to get the desired rotation.

So thanks For your Answer

I know We must have 3 Volume

IS it correct:

I Can Import My geometry ( AFT & FOR Propeller + FarFeild ) And Separate it (By Turn Off the block Of Them) And Have 3 different type Of Mesh (For Example Structure & UnStructure Mesh) And Save each of Volume separately;
And export them to Fluent ?


My Question is it When I Have different Type Of Mesh (both Have different Size Function Or Different type of Mesh)
Is it not problem(Error) On Interface of Volumes in Solvers (ex:Fluent or CFX)?

And How to Coordinate them on solvers ?( How Export them on ICEM or HOW to Import them to Solvers To Coupled Them (That The 3 Volume Mach Excatly )??

PSYMN July 17, 2013 17:32

CFX allows you to import multiple separate files, but Fluent wants you to import just one file with multiple zones.

This just means you need to combine them in another tool first. For instance, you could mesh all three files in separate ICEM CFD projects and then open a 4th project (or load them into TGrid or where ever) and combine them and then output a final combined project.

You could have 3 geometry sub models, or just a single geometry and only zoom in and mesh partial regions...

When you load 3 *.uns or *.msh files into ICEM CFD, they are concatenated. This means they are in one file, but nodes are not merged, etc. Perfect for output to Fluent.

Best regards,

Simon

Far July 18, 2013 02:02

did you check this one?

http://www.cfd-online.com/Forums/ans...ler-blade.html

ahmadreza July 18, 2013 10:43

Quote:

Originally Posted by Far (Post 440451)


Hi
Thanks For Your Attention

I try It but a quality of mesh Was Not Good

Are You have Other Suggestion ??

Far July 18, 2013 10:54

Simon has already given very accurate suggestions.

Take one blade from both propeller and make mesh for them. You can get all blades in CFX-Pre and which can imported later to Fluent or you can copy mesh or blocking to get mesh for all blades as simon told you.

You need three volume (one for each rotating domain and one for farfield) as already indicated by Simon. But make sure that you have separate faces in fairfield volume for each blade volume. They should exactly match in axial extent.

Are you willing to share your project?

Far July 18, 2013 13:10

http://www.forwind.de/sowe/Site/Prog...E2013_Dose.pdf


In first propeller you have round leading and trailing edge and also blade tip.

In 2nd propeller sharp leading edge and trailing edge with blade tip.

In both propellers you straight periodic boundaries, so make them according to blade shape.

ahmadreza July 18, 2013 13:18

Quote:

Originally Posted by Far (Post 440615)
http://www.forwind.de/sowe/Site/Prog...E2013_Dose.pdf


In first propeller you have round leading and trailing edge and also blade tip.

In 2nd propeller sharp leading edge and trailing edge with blade tip.

In both propellers you straight periodic boundaries, so make them according to blade shape.


Thanks

Can I Have One Request..

Please Send Me First Propeller With Block that you think it is good (block with boundary condition and Rotate)
(please Email to me)

handsome November 11, 2013 23:44

sharp edge volume orientation error
 
Hi
Question
I want to mesh the body which have a sharp edge .but when the mesh was completed I see some element like the picture below. Some of this bad element is belong the body and when I invisible the body the disappear .but some of the bad element can be seen even when I invisible all parts of the body and when I use check mesh the put in added faces.
When I mesh the body by mesh size of 0.02 everything is ok but when reduce the max mesh size to 0.01 the bad element will appear.
(I have reduced the tri tolerance and edge criterion number )
http://img.majidonline.com/thumb/299...1020130561.jpg


All times are GMT -4. The time now is 11:22.