Thin Plate Moves Between 2 Parallel Surfaces

ordinary · January 2, 2018, 17:12

Hello everyone,

Here is the original question:

problem.jpg

There are two regions. They are same(h1=h2). So this problem might be considered as 'semi lid-driven cavity' problem. But only one difference: Inlet and outlet are not wall. Therefore I encounter reverse flow error.
I drew a rectangle and let the south boundary to moving wall(absolute 5 m/s), north boundary to wall, east boundary to pressure outlet with 0 and west boundary to velocity-inlet with 0 m/s.
Eventually I failed.
Any idea how to solve ?

Regards.

DEd · January 2, 2018, 18:03

Why did you use a wall for the west and the east BC? Wheren't these suppose to be inlets?

But I see your problem. What to use as an inlet BC. I would expand my domain until I can define a BC.

ordinary · January 2, 2018, 21:23

Thank you for your attention.

I corrected boundary conditions. My aim is to find reaction force of moving plate due to drag by viscosity. But I coudn't set up the problem.

AlexanderZ · January 3, 2018, 00:23

I think you may use symmetry for this case

best regards

DEd · January 3, 2018, 05:10

I don't see the place for symmetry (except at the moving south moving wall but i guess you omit the south domain right) wall but maybe I am missing something. Could you post a picture woth your fluent case and also the failed/perhaps erronous solution?

ordinary · January 3, 2018, 06:38

Thank you everyone for your attention. I post some images to clear misunderstandings:

Here is the problem:

scheme.png

Dynamic viscosity is 0.9 Ns/m^2 and density is 1000 kg/m^3.

Thanks to the horizontal symmetry, I only drew a rectangle as shown:

simple.png

I've done the mesh part.

If I choose 2D Space: Axisymmetric, initialization gives me Error: floating point exception Error Object: #f

If I choose 2D Space: Planar, initialization is OK. But it doesn't converge. It always gives me famous reversed flow error.

Thank you again.

Best regards.

DEd · January 3, 2018, 10:11

Ok, then I understand what symmetry you were talking about.

So what BC did you use for inlet BC? I think with this domain two (static) pressure outlet is most suitable but I suspect it will warn for "backflow" at least at the inlet (west BC).

Otherwise you might try a different domain. Someting like this perhaps where you don't need any inlet och outlets (like the lid driven cavity case).

Ps. Even non converged cases can give clues to problems so they can be good to look at/post

Ps2. 2d planar is the way to go. Axi symmetric would mean your plate is a string/line

ordinary · January 3, 2018, 11:00

Now I'm trying periodic BC for left and right side of a simple rectangle(2m width, 0.04m height). Bottom of the rectangle is moving wall with 10 m/s. Top part is stationary wall of course.

For mesh part, cells grows as they goes up. Cells near moving wall is smaller cells. Cell sizes only changes vertically, not horizontally.

But it is not converging:

notconverge.png

Geometry for new case:

geometry.PNG

DEd · January 3, 2018, 16:38

Your case seemed simple enough so I tried setting it up. I have no problems with getting convergence (also checked convergence of integral of wall shear stress on the moving wall).

BCs
Inlet: Pressure Outlet (0 Pa)
Outlet: Pressure Outlet (0 Pa)
Moving wall: 10 m/s

Mesh:

ordinary · January 3, 2018, 17:48

Periodic BCs works perfect. Prossure outlet BCs gives wrong solution. (totally irrelevant to analytical solution)

Regards
Emre

DEd · January 6, 2018, 05:54

Quote:

Originally Posted by ordinary

Periodic BCs works perfect. Prossure outlet BCs gives wrong solution. (totally irrelevant to analytical solution)

Regards
Emre

What do you mean with "wrong solution"? It gives another solution.

Certainly, if it was specified that it was many plates after each other periodic BC might be suitable but for a single moving wall as described the solution will show asymmetry (more developed flow on the outlet).

ordinary · January 6, 2018, 06:01

Hello. I mean wrong results. It is supposed to have 225 Pa shear stress. Pressure outlet BC gives 467 Pa. But periodic BC calculates exactly 225 Pa.

DEd · January 6, 2018, 06:29

Can you publish the original description of the problem?

ordinary · January 6, 2018, 06:56

Here is the solution what satisfied my analytical calculations:

Periodic BC with simple rectangle.

Shear stress is -225 Pa.

1.jpg

Here I tried to calculate force(Multiplied by 2, because there is also another same shear stress acting on moving wall which is below the moving wall)

2.PNG

Total force is -450 N which is 100% correct.

3.PNG

BTW, ANSYS zipped case&data files exceeds attachment limitations.
4.jpg

DEd · January 6, 2018, 07:07

Thanks for adding the full description of the problem in your original thread.

However, I maintain that the way the problem is described a periodic BC is a stretch as there will be entry effects. If they said infinitly long and asked after force per meter on the other hand a periodic BC is a fit. Text book makers aren't perfect

ordinary · January 6, 2018, 07:15

Actually there is no "infinite" word in the text. So it's a bit confusing. But I assume it as infinite Couette flow. Because analytical solution is,

Tau1 = 0.9*10/0.04 = 225 Pa
Tau2 = 0.9*10/0.04 = 225 Pa
F = (Tau1 + Tau2) * Area = 2*225*(2*0.5)= -450 Pa due to viscosity.

Thank you again for your attention.
Regards.

January 2, 2018, 17:12	Thin Plate Moves Between 2 Parallel Surfaces	#1
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Hello everyone, Here is the original question: problem.jpg There are two regions. They are same(h1=h2). So this problem might be considered as 'semi lid-driven cavity' problem. But only one difference: Inlet and outlet are not wall. Therefore I encounter reverse flow error. I drew a rectangle and let the south boundary to moving wall(absolute 5 m/s), north boundary to wall, east boundary to pressure outlet with 0 and west boundary to velocity-inlet with 0 m/s. Eventually I failed. Any idea how to solve ? Regards. Last edited by ordinary; January 6, 2018 at 06:41. Reason: Final edit - Original problem

January 3, 2018, 11:00		#8
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Now I'm trying periodic BC for left and right side of a simple rectangle(2m width, 0.04m height). Bottom of the rectangle is moving wall with 10 m/s. Top part is stationary wall of course. For mesh part, cells grows as they goes up. Cells near moving wall is smaller cells. Cell sizes only changes vertically, not horizontally. But it is not converging: notconverge.png Geometry for new case: geometry.PNG Last edited by ordinary; January 6, 2018 at 06:42. Reason: I have no idea why this didn't work. But periodic BC works.

January 3, 2018, 16:38		#9
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	Your case seemed simple enough so I tried setting it up. I have no problems with getting convergence (also checked convergence of integral of wall shear stress on the moving wall). BCs Inlet: Pressure Outlet (0 Pa) Outlet: Pressure Outlet (0 Pa) Moving wall: 10 m/s Mesh: ordinary likes this.

January 3, 2018, 17:48		#10
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Periodic BCs works perfect. Prossure outlet BCs gives wrong solution. (totally irrelevant to analytical solution) Regards Emre Last edited by ordinary; January 5, 2018 at 10:19. Reason: Sorry for calculation errors.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2way FSI for thin plate, mapping problem	behest	ANSYS	2	August 16, 2016 10:19
[GAMBIT] Making hundereds of small hole in a thin plate using Gambit	irfanlhr	ANSYS Meshing & Geometry	2	October 23, 2012 07:05
Parallel Plate Rheometer vs Cone & Plate	beginner	Main CFD Forum	0	March 19, 2003 11:49
iso surfaces in parallel	Frank Muldoon	Main CFD Forum	2	July 11, 2002 13:17
free convection between 2 infinite parallel plate	runny craven	Main CFD Forum	2	September 19, 2001 05:45

January 2, 2018, 18:03		#2
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	Why did you use a wall for the west and the east BC? Wheren't these suppose to be inlets? But I see your problem. What to use as an inlet BC. I would expand my domain until I can define a BC.

January 2, 2018, 21:23		#3
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Thank you for your attention. I corrected boundary conditions. My aim is to find reaction force of moving plate due to drag by viscosity. But I coudn't set up the problem.

January 3, 2018, 00:23		#4
AlexanderZ Senior Member Alexander Join Date: Apr 2013 Posts: 2,363 Rep Power: 34	I think you may use symmetry for this case best regards

January 3, 2018, 05:10		#5
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	I don't see the place for symmetry (except at the moving south moving wall but i guess you omit the south domain right) wall but maybe I am missing something. Could you post a picture woth your fluent case and also the failed/perhaps erronous solution?

January 3, 2018, 06:38		#6
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Thank you everyone for your attention. I post some images to clear misunderstandings: Here is the problem: scheme.png Dynamic viscosity is 0.9 Ns/m^2 and density is 1000 kg/m^3. Thanks to the horizontal symmetry, I only drew a rectangle as shown: simple.png I've done the mesh part. If I choose 2D Space: Axisymmetric, initialization gives me Error: floating point exception Error Object: #f If I choose 2D Space: Planar, initialization is OK. But it doesn't converge. It always gives me famous reversed flow error. Thank you again. Best regards.

January 3, 2018, 10:11		#7
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	Ok, then I understand what symmetry you were talking about. So what BC did you use for inlet BC? I think with this domain two (static) pressure outlet is most suitable but I suspect it will warn for "backflow" at least at the inlet (west BC). Otherwise you might try a different domain. Someting like this perhaps where you don't need any inlet och outlets (like the lid driven cavity case). Ps. Even non converged cases can give clues to problems so they can be good to look at/post Ps2. 2d planar is the way to go. Axi symmetric would mean your plate is a string/line

January 6, 2018, 06:01		#12
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Hello. I mean wrong results. It is supposed to have 225 Pa shear stress. Pressure outlet BC gives 467 Pa. But periodic BC calculates exactly 225 Pa.

January 6, 2018, 06:29		#13
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	Can you publish the original description of the problem?

January 6, 2018, 06:56		#14
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Here is the solution what satisfied my analytical calculations: Periodic BC with simple rectangle. Shear stress is -225 Pa. 1.jpg Here I tried to calculate force(Multiplied by 2, because there is also another same shear stress acting on moving wall which is below the moving wall) 2.PNG Total force is -450 N which is 100% correct. 3.PNG BTW, ANSYS zipped case&data files exceeds attachment limitations. 4.jpg

January 6, 2018, 07:07		#15
DEd Member Daniel Edebro Join Date: Feb 2016 Location: Gothenburg Posts: 41 Rep Power: 10	Thanks for adding the full description of the problem in your original thread. However, I maintain that the way the problem is described a periodic BC is a stretch as there will be entry effects. If they said infinitly long and asked after force per meter on the other hand a periodic BC is a fit. Text book makers aren't perfect

January 6, 2018, 07:15		#16
ordinary Member Emre Join Date: Nov 2015 Location: Izmir, Turkey Posts: 97 Rep Power: 10	Actually there is no "infinite" word in the text. So it's a bit confusing. But I assume it as infinite Couette flow. Because analytical solution is, Tau1 = 0.910/0.04 = 225 Pa Tau2 = 0.910/0.04 = 225 Pa F = (Tau1 + Tau2) * Area = 2225(2*0.5)= -450 Pa due to viscosity. Thank you again for your attention. Regards.