mesh stiffness of FSI

colopolo · September 20, 2011, 03:13

Dear All,

I am trying to run vibration of plate using fsi (to find critical flutter velocity)

I have a big problem about folded mesh. I have been searched this problem in here and google but I cannot solve it.

I did control mesh stiffness of increase near boundary and increase near small volume. I used 1 or 10.

But every running case has folded mesh.

I really have no idea how to control foled mesh.

Please welcome any ideas for this problem

Thanks in advance

mvoss · September 20, 2011, 04:06

hey,

you need to create an expression for the stiffening around the corners. smth. like (sqrt((x-x0)^2+(y-y0)^2))^4 with x0/y0 being the coordinates of the current location of one corner e.g. (it´s a circle function to the power of 4 to make it really hard at the pref. point).

Create an expression for every point or even for smth. like a line function and overlay them by summation in a final expression for the stiffness.

The tricky thing is to update the x0/y0 every timestep. There is a solution for this ... smth. dealing with monitorpoints.... i don´t remember exactly.

neewbie

colopolo · September 20, 2011, 04:10

Quote:

Originally Posted by neewbie

hey,

you need to create an expression for the stiffening around the corners. smth. like (sqrt((x-x0)^2+(y-y0)^2))^4 with x0/y0 being the coordinates of the current location of one corner e.g. (it´s a circle function to the power of 4 to make it really hard at the pref. point).

Create an expression for every point or even for smth. like a line function and overlay them by summation in a final expression for the stiffness.

The tricky thing is to update the x0/y0 every timestep. There is a solution for this ... smth. dealing with monitorpoints.... i don´t remember exactly.

neewbie

Thanks you your response. I am not familiar with expression formula.
Do I just put the above equation in the mesh stiffness term of CFX?
Could you tell me a little bit detail about how to create it?

This is wind-induced vibration problem so how to define x0 and y0?

Sorry to bother you.

Thanks.

mvoss · September 20, 2011, 04:14

take a look at the tutorials ... creating the expressions is straight forward.
create an expression stiff1, stiff2 with the app. expressions for x1,y1,x2,y2 and combine them to let´s say: stiffness=stiff1+stiff2

colopolo · September 20, 2011, 04:28

Quote:

Originally Posted by neewbie

take a look at the tutorials ... creating the expressions is straight forward.
create an expression stiff1, stiff2 with the app. expressions for x1,y1,x2,y2 and combine them to let´s say: stiffness=stiff1+stiff2

Thanks.. I will try to do that. I really appreciate your help

mvoss · September 20, 2011, 06:18

Quote:

Originally Posted by neewbie

... (sqrt((x-x0)^2+(y-y0)^2))^4

it has to be $\sqrt{(x-x_{0})^{2}+(y-y_{0})^{2}}^{-4}$

colopolo · September 20, 2011, 18:03

Quote:

Originally Posted by neewbie

it has to be $\sqrt{(x-x_{0})^{2}+(y-y_{0})^{2}}^{-4}$

Very appreciate it. I tried to make CEL expression but I do not have no idea how to define x0 and y0.
Does this x0 and y0 indicate the coordinate of edge of plate?
if you help me more,. thank you very much.

colopolo · September 20, 2011, 18:47

Quote:

Originally Posted by colopolo

Very appreciate it. I tried to make CEL expression but I do not have no idea how to define x0 and y0.
Does this x0 and y0 indicate the coordinate of edge of plate?
if you help me more,. thank you very much.

I changed more simple tetra mesh and attached a mesh image just before error.

I supposed the error occur near the plate but, in my case, large mesh deformation occur between inlet and plate.

Does anyone has any ideas?

I am still trying to make CEL which you mentioned.

mvoss · September 21, 2011, 09:28

just guessing... but wouldn´t smth. like C*f(x) for the stiffness solve this similar to the idea above? Isn´t there a "stiff near bc" option?

colopolo · September 21, 2011, 19:43

Quote:

Originally Posted by neewbie

just guessing... but wouldn´t smth. like C*f(x) for the stiffness solve this similar to the idea above? Isn´t there a "stiff near bc" option?

Thanks for ur response. There is "stiff near bc'' option. I also controlled 'stiff near bc' as 1, 10, and 100. but nothing changed.

Darren Leong · September 22, 2011, 03:52

Leave the exponent value at 10. Looking at your pic, the mesh in the farfield is skewing. Try enlarging the domain or have a coarser mesh in the farfield which will allow more deformation overall

colopolo · September 22, 2011, 04:15

Quote:

Originally Posted by Darren Leong

Leave the exponent value at 10. Looking at your pic, the mesh in the farfield is skewing. Try enlarging the domain or have a coarser mesh in the farfield which will allow more deformation overall

Thanks !! I will try to do that and leave the result in this thread..

Very appreciate it

Also, I have another problem. the plate moves at x -direction even if the plate is constrained in x-dir.

I onle let it free moving at y-dir and rotating at z-dir but it moves also x-dir.

colopolo · September 22, 2011, 21:50

Quote:

Originally Posted by Darren Leong

Leave the exponent value at 10. Looking at your pic, the mesh in the farfield is skewing. Try enlarging the domain or have a coarser mesh in the farfield which will allow more deformation overall

It looks ok. but Force can not converge after some time stpes.
Force is increasing more and more.. how to control force convergence?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
FSI Mesh stiffness	Andrew.Coughtrie	CFX	3	July 29, 2011 03:50
[snappyHexMesh] external flow with snappyHexMesh	chelvistero	OpenFOAM Meshing & Mesh Conversion	11	January 15, 2010 19:43
Convergence moving mesh	lr103476	OpenFOAM Running, Solving & CFD	30	November 19, 2007 14:09
Mesh Stiffness Option for Mesh Deformation	Ste_Lakey	CFX	3	January 19, 2006 16:33
How to control Minximum mesh space?	hung	FLUENT	7	April 18, 2005 09:38

September 20, 2011, 04:06		#2
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	hey, you need to create an expression for the stiffening around the corners. smth. like (sqrt((x-x0)^2+(y-y0)^2))^4 with x0/y0 being the coordinates of the current location of one corner e.g. (it´s a circle function to the power of 4 to make it really hard at the pref. point). Create an expression for every point or even for smth. like a line function and overlay them by summation in a final expression for the stiffness. The tricky thing is to update the x0/y0 every timestep. There is a solution for this ... smth. dealing with monitorpoints.... i don´t remember exactly. neewbie

September 20, 2011, 04:14		#4
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	take a look at the tutorials ... creating the expressions is straight forward. create an expression stiff1, stiff2 with the app. expressions for x1,y1,x2,y2 and combine them to let´s say: stiffness=stiff1+stiff2

September 21, 2011, 09:28		#9
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	just guessing... but wouldn´t smth. like C*f(x) for the stiffness solve this similar to the idea above? Isn´t there a "stiff near bc" option?

September 22, 2011, 03:52		#11
Darren Leong Member Darren Leong Join Date: Dec 2010 Posts: 64 Rep Power: 15	Leave the exponent value at 10. Looking at your pic, the mesh in the farfield is skewing. Try enlarging the domain or have a coarser mesh in the farfield which will allow more deformation overall