Problem With Deforming Mesh

bertozzi_marco · April 27, 2015, 09:58

Hallo everybody
I've a problem with deforming mesh I've ever seen,
There is a perfect hex mesh made with ICEM, and a simply moving wall.
Happens that the mesh when the simulation go on, become less parallel
as the time go forward.
To better explain the problem there is an animation of the mesh and a particular of the distorsion of the mesh after a lot of iteration.
You can find files on the link below.
Why the mesh lose parallelism during simulation ?

https://www.dropbox.com/sh/zz5x4soel...Z1y-gDoda?dl=0

singer1812 · April 27, 2015, 12:07

I am not able to go into drop box, but what you describe can be done easily.

You need to write a CEL script to update mesh deformations precisely. Instead of moving just the boundary, create a subdomain and deform the subdomain.

Your deformation will be 1 at the boundary, 0 at the fixed wall, and linear function in between.

You should be able to search in this forum for more details on this approach.

bertozzi_marco · April 27, 2015, 13:07

Thankyou Singer,
you are rigth and I will try your suggestion, now I've bypassed the
problem adding a remesh command into CFX, but I don't understand
why the top geometry that move in the same way mantain perfect parallelism between elements, and the middle geometry deform in such way the elements.
I don't see any difference.... does anyone know the reason?

bertozzi_marco · April 28, 2015, 11:36

Sorry Singer,
I've looked for your suggestion in this forum and I haven't found anything,
plese can you expalin little bit when you say to drive precisely the mesh with cel scrip ? Pheraps you mean junction box routine ?

singer1812 · April 29, 2015, 09:56

Specify your subdomain and expressions. For example the below is a subdomain that "squashes up in the Global X direction. x2 is the stationary wall location (it will move at 0*movefront) and x1 is the moving wall side (it will move at 1*movefront). In between meshes will get linearly "squashed" (zero to one*movefront based on X location) in between. Travel(t) is a tabular function of my travel versus time (you can make these directly in Pre).

EXPRESSIONS:
movefront = (Travel(t)-Travel(tstart))*min(1,max(0, (x2-x)/(x2-x1)))
tstart = .0 [s]
x1 = minVal(xGlobal)@Sub_Front
x2 = maxVal(xGlobal)@Sub_Front
END

SUBDOMAIN: Sub_Front
Coord Frame = Coord 0
Location = FRONT_AIR
MESH MOTION:
Option = Specified Displacement
DISPLACEMENT:
Displacement X Component = movefront
Displacement Y Component = 0 [m]
Displacement Z Component = 0 [m]
Option = Cartesian Components
END
END
END

bertozzi_marco · April 29, 2015, 10:36

Thankyou singer,
I'hadn't seen the meaning of your suggestion, now I've done and
everything work perfect thankyou verymuch

Now if you are so kind, I'd ask any more thing: I've the results
of a transient analisys for a period of 10sec.
I'd like to calculate the mechanical work of the mechanism during one
cycle. I've the monitor out file of the pressure and the monitor of the
positions of the moving walls. With excel is easy to combine the results
and obtain the work / cycle, but it'snt usefull for my final goal.

I whant to do a geometric optimizaton of the model and I've made
a series of design points. I'd like to know if there is a way to interpolate the
results of a transient analisys in order to obtain a unique value to be used for
the goal of an optimizaton study.

April 27, 2015, 09:58	Problem With Deforming Mesh	#1
bertozzi_marco Member Marco Join Date: Jul 2013 Location: Italy Posts: 36 Rep Power: 13	Hallo everybody I've a problem with deforming mesh I've ever seen, There is a perfect hex mesh made with ICEM, and a simply moving wall. Happens that the mesh when the simulation go on, become less parallel as the time go forward. To better explain the problem there is an animation of the mesh and a particular of the distorsion of the mesh after a lot of iteration. You can find files on the link below. Why the mesh lose parallelism during simulation ? https://www.dropbox.com/sh/zz5x4soel...Z1y-gDoda?dl=0

April 29, 2015, 09:56		#5
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 21	Specify your subdomain and expressions. For example the below is a subdomain that "squashes up in the Global X direction. x2 is the stationary wall location (it will move at 0movefront) and x1 is the moving wall side (it will move at 1movefront). In between meshes will get linearly "squashed" (zero to onemovefront based on X location) in between. Travel(t) is a tabular function of my travel versus time (you can make these directly in Pre). EXPRESSIONS: movefront = (Travel(t)-Travel(tstart))min(1,max(0, (x2-x)/(x2-x1))) tstart = .0 [s] x1 = minVal(xGlobal)@Sub_Front x2 = maxVal(xGlobal)@Sub_Front END SUBDOMAIN: Sub_Front Coord Frame = Coord 0 Location = FRONT_AIR MESH MOTION: Option = Specified Displacement DISPLACEMENT: Displacement X Component = movefront Displacement Y Component = 0 [m] Displacement Z Component = 0 [m] Option = Cartesian Components END END END bertozzi_marco likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] No layers in a small gap	bobburnquist	OpenFOAM Meshing & Mesh Conversion	6	August 26, 2015 10:38
Overset MESH problem DFBI 6DOF	Ale85	STAR-CCM+	0	October 1, 2013 13:25
CFX Mesh Deformation problem	Silmaril	CFX	7	October 19, 2010 11:00
fluent add additional zones for the mesh file	SSL	FLUENT	2	January 26, 2008 12:55
How to control Minximum mesh space?	hung	FLUENT	7	April 18, 2005 10:38

April 27, 2015, 12:07		#2
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 21	I am not able to go into drop box, but what you describe can be done easily. You need to write a CEL script to update mesh deformations precisely. Instead of moving just the boundary, create a subdomain and deform the subdomain. Your deformation will be 1 at the boundary, 0 at the fixed wall, and linear function in between. You should be able to search in this forum for more details on this approach.

April 27, 2015, 13:07		#3
bertozzi_marco Member Marco Join Date: Jul 2013 Location: Italy Posts: 36 Rep Power: 13	Thankyou Singer, you are rigth and I will try your suggestion, now I've bypassed the problem adding a remesh command into CFX, but I don't understand why the top geometry that move in the same way mantain perfect parallelism between elements, and the middle geometry deform in such way the elements. I don't see any difference.... does anyone know the reason?

April 28, 2015, 11:36		#4
bertozzi_marco Member Marco Join Date: Jul 2013 Location: Italy Posts: 36 Rep Power: 13	Sorry Singer, I've looked for your suggestion in this forum and I haven't found anything, plese can you expalin little bit when you say to drive precisely the mesh with cel scrip ? Pheraps you mean junction box routine ?

April 29, 2015, 10:36		#6
bertozzi_marco Member Marco Join Date: Jul 2013 Location: Italy Posts: 36 Rep Power: 13	Thankyou singer, I'hadn't seen the meaning of your suggestion, now I've done and everything work perfect thankyou verymuch Now if you are so kind, I'd ask any more thing: I've the results of a transient analisys for a period of 10sec. I'd like to calculate the mechanical work of the mechanism during one cycle. I've the monitor out file of the pressure and the monitor of the positions of the moving walls. With excel is easy to combine the results and obtain the work / cycle, but it'snt usefull for my final goal. I whant to do a geometric optimizaton of the model and I've made a series of design points. I'd like to know if there is a way to interpolate the results of a transient analisys in order to obtain a unique value to be used for the goal of an optimizaton study.