What is the diffrence between Steady and Transient 2 Way FSI simulation?

michellemay1993 · June 27, 2016, 05:42

Hi,

Can anyone explain what the difference between steady and transient 2 way FSI simulation is?

ghorrocks · June 27, 2016, 07:57

2 way FSI means the fluid affects the structure, and the structural movement affects the flow. Transient means the result changes with time, steady state means the result is constant with time (ie , it reaches an equilibrium position and just stays there).

michellemay1993 · June 27, 2016, 08:05

Thank you for your reply, ghorrocks.

But I am not really understand. Is there any example?

ghorrocks · June 27, 2016, 08:07

Steady state means it reaches equilibrium. That is, it is still. So the fluid field and the structural thing does not move.

Transient means it is moving.

michellemay1993 · June 27, 2016, 08:37

Therefore, in a peristaltic pump simulation, I should use transient type analysis?

Furthermore, what is the proper boundary condition for a peristaltic pump simulation?

ghorrocks · June 27, 2016, 08:56

Most peristaltic pumps I know of are in constant motion. So that must be a transient analysis.

The boundary conditions should be selected to match the flow you are modelling. That depends on the system this pump is part of. So we cannot answer this question until you describe how the pump is connected to the fluid system around it.

michellemay1993 · June 27, 2016, 12:39

Actually I am new in this. Therefore, in the simulation, I replaced the piezoelectric bimorph with forces and the mass flow rate should reach a certain value without back pressure.

Here are my boundary condition:
Inlet and Outlet: opening pres. and Dirn, relative pressure: 0 Pa

Wall: Ansys Multifield

However, I have been getting the same error all the time ever I have done some changes to the boundary condition.

Can you help me?

ghorrocks · June 27, 2016, 20:46

Please attach an image of your geometry and mesh, and you output file containing the error message.

michellemay1993 · June 27, 2016, 23:33

Here is my full geometry.

Here is my fluid geometry.

Here is the mesh generated by ICEM CFD.

I have copied the outfile details and pasted into a text file.

Thank you very much for helping me in this.

ghorrocks · June 28, 2016, 02:23

Do you know the motion of the pump in advance? Does the fluid affect the shape the walls take?

michellemay1993 · June 28, 2016, 02:39

The pump has a traveling wave on the top surface of the pump which will push the fluid out like peristaltic pump.

I'm not really sure the fluid affects the shape of wall...

ghorrocks · June 28, 2016, 02:51

So I would start by assuming that the fluid does not affect the shape. This means the simulation is no longer an FSI simulation but is a moving mesh which is much simpler. So I recommend you do a moving mesh simulation on this and get that working before you do FSI. FSI is MUCH more complex, and you have no hope of getting FSI to work if you can't get the simpler moving mesh working.

michellemay1993 · June 28, 2016, 03:55

Any example that I can follow?

ghorrocks · June 28, 2016, 06:39

There are a few moving mesh tutorial examples which come with CFX. Have a look in the CFX documentation for the tutorials.

michellemay1993 · June 28, 2016, 07:48

From the tutorial, the only possible domain motion available is rotation. So, how to implement translational motion? Do you have any example with translational mesh motion?

ghorrocks · June 28, 2016, 09:24

You are talking about rotating frames of reference. You cannot use that for translational motion, this is a different thing. I am referring to moving mesh - it is used in the Buoy model with rigid body solver. You will not need the rigid body stuff, but the mesh motion is used in that example.

michellemay1993 · June 28, 2016, 11:18

Is it possible to use one way fsi?

michellemay1993 · June 28, 2016, 19:41

When using 2 way fsi in this micropump, it actually work with bigger time step size like 0.0025s but generated incorrect velocity profile. Since the forces are cyclic forces(cosine wave) and it work with frequency of 1437 Hz therefore I reduce the time step size to 1/(20*1437) s but turns out cannot run at all with any boundary condition. What should I do?

ghorrocks · June 28, 2016, 20:52

FSI is complex and if you can't get a moving mesh simulation to run then you have no hope in getting FSI to run. Note that FSI uses moving mesh to model the motions. So you really should start with a moving mesh simulation for this case.

Forget the FSI for now.

michellemay1993 · July 13, 2016, 12:42

Can I know how to create a wave like mesh motion on a wall?

wave motion like ----> Y=4A*cos(2*pi*x/lambda-wt)

June 27, 2016, 05:42	What is the diffrence between Steady and Transient 2 Way FSI simulation?	#1
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Hi, Can anyone explain what the difference between steady and transient 2 way FSI simulation is?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2way FSi, Initialize with steady solution, Fluent, Transient Sturcural, System Coupli	mmkkeshavarzi	FLUENT	1	June 28, 2024 08:03
Transient simulation -> Steady state solution	saisanthoshm88	CFX	13	April 10, 2024 05:43
The difference between steady state and transient	JuPa	CFX	36	December 9, 2019 23:50
Steady state simulation with transient partilcle tracking	mali28	FLUENT	2	February 7, 2013 15:25
Ramping Boundary Condition for Steady FSI Simulation	ashtonJ	CFX	3	August 21, 2012 04:18

June 27, 2016, 07:57		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	2 way FSI means the fluid affects the structure, and the structural movement affects the flow. Transient means the result changes with time, steady state means the result is constant with time (ie , it reaches an equilibrium position and just stays there).

June 27, 2016, 08:05		#3
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Thank you for your reply, ghorrocks. But I am not really understand. Is there any example?

June 27, 2016, 08:07		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Steady state means it reaches equilibrium. That is, it is still. So the fluid field and the structural thing does not move. Transient means it is moving.

June 27, 2016, 08:37		#5
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Therefore, in a peristaltic pump simulation, I should use transient type analysis? Furthermore, what is the proper boundary condition for a peristaltic pump simulation?

June 27, 2016, 08:56		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Most peristaltic pumps I know of are in constant motion. So that must be a transient analysis. The boundary conditions should be selected to match the flow you are modelling. That depends on the system this pump is part of. So we cannot answer this question until you describe how the pump is connected to the fluid system around it.

June 27, 2016, 12:39		#7
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Actually I am new in this. Therefore, in the simulation, I replaced the piezoelectric bimorph with forces and the mass flow rate should reach a certain value without back pressure. Here are my boundary condition: Inlet and Outlet: opening pres. and Dirn, relative pressure: 0 Pa Wall: Ansys Multifield However, I have been getting the same error all the time ever I have done some changes to the boundary condition. Can you help me?

June 27, 2016, 20:46		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Please attach an image of your geometry and mesh, and you output file containing the error message.

June 27, 2016, 23:33		#9
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Here is my full geometry. Here is my fluid geometry. Here is the mesh generated by ICEM CFD. I have copied the outfile details and pasted into a text file. Thank you very much for helping me in this.

June 28, 2016, 02:23		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Do you know the motion of the pump in advance? Does the fluid affect the shape the walls take?

June 28, 2016, 02:39		#11
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	The pump has a traveling wave on the top surface of the pump which will push the fluid out like peristaltic pump. I'm not really sure the fluid affects the shape of wall...

June 28, 2016, 02:51		#12
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	So I would start by assuming that the fluid does not affect the shape. This means the simulation is no longer an FSI simulation but is a moving mesh which is much simpler. So I recommend you do a moving mesh simulation on this and get that working before you do FSI. FSI is MUCH more complex, and you have no hope of getting FSI to work if you can't get the simpler moving mesh working.

June 28, 2016, 03:55		#13
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Any example that I can follow?

June 28, 2016, 06:39		#14
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	There are a few moving mesh tutorial examples which come with CFX. Have a look in the CFX documentation for the tutorials.

June 28, 2016, 07:48		#15
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	From the tutorial, the only possible domain motion available is rotation. So, how to implement translational motion? Do you have any example with translational mesh motion?

June 28, 2016, 09:24		#16
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	You are talking about rotating frames of reference. You cannot use that for translational motion, this is a different thing. I am referring to moving mesh - it is used in the Buoy model with rigid body solver. You will not need the rigid body stuff, but the mesh motion is used in that example.

June 28, 2016, 11:18		#17
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Is it possible to use one way fsi?

June 28, 2016, 19:41		#18
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	When using 2 way fsi in this micropump, it actually work with bigger time step size like 0.0025s but generated incorrect velocity profile. Since the forces are cyclic forces(cosine wave) and it work with frequency of 1437 Hz therefore I reduce the time step size to 1/(20*1437) s but turns out cannot run at all with any boundary condition. What should I do?

June 28, 2016, 20:52		#19
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	FSI is complex and if you can't get a moving mesh simulation to run then you have no hope in getting FSI to run. Note that FSI uses moving mesh to model the motions. So you really should start with a moving mesh simulation for this case. Forget the FSI for now.

July 13, 2016, 12:42		#20
michellemay1993 New Member Join Date: Jun 2016 Posts: 25 Rep Power: 10	Can I know how to create a wave like mesh motion on a wall? wave motion like ----> Y=4Acos(2pi*x/lambda-wt)