CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > ANSYS > CFX

Possibilities to influence Simulation in CFX

Register Blogs Community New Posts Updated Threads Search

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   October 30, 2014, 05:28
Default
  #21
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
I found a paper from Ansys with Information about transient simulation in CFX. There by I have realised the meaning of some values.

To take some experience I will start a transient Simulation and if there are any question after the run I will report another post.


Unfortunately the file size of the paper exceed the maximum file size, so I can't attache for other user.
biro is offline   Reply With Quote

Old   October 30, 2014, 18:04
Default
  #22
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
I have set up my first transient simulation. For Time Duration I set 10s (but now I think the value is to high to reach).

There by for me arise three further question:
1. Whats the difference between "Analyse Type" --> "Timestep Adaption" --> "Target Max (Min) Loops" and "Solver Control" --> "Convergence Control" --> "Max (Min) Coeff Loops.
I set Target Max Loops = 5; Target Min Loops = 3 and for Min Coeff Loops (=1) and Max Coeff Loops (=10) I don't change the default values.
Is this right or did I misunderstood the post of ghorocks:
Quote:
The recommended approach is to use adaptive timesteps. Set it to home in on 3-5 coeff loops per iteration, and make the minimum and maximum allowable time step wide enough that it will never reach it.
2. For transient Results (Output Control) I set a timestep Intervall of 8 ? Is this a good value ? I can't imagine, which value I have to choose.

3. What will I see in the Transient Result Files.


Thank you for your help!

Greets
biro is offline   Reply With Quote

Old   October 30, 2014, 18:07
Default
  #23
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
1. Those values look OK
2. Note that as the time step size will vary the time spacing of the results will vary. But you appear to only be interested in the final result so you might not need transient results files at all.
3. snapshots of the flow at that time step.
ghorrocks is offline   Reply With Quote

Old   October 30, 2014, 18:14
Default
  #24
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
thank you for the fast reply.

For my opinion in the result file there are no transient Information and with the transient files that all "build" my transient simulation.
But now I think thats wrong.


I will take a look on my simualtion the next days, I think there will be more question
biro is offline   Reply With Quote

Old   November 3, 2014, 07:29
Default
  #25
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
The trend of my transient simulation looks very good (for me).
I attache the progress of the residuals and imbalances, pressure at the interesting Monitorpoints and Reynoldsnumber at the interesting region.

With this results:
Can I say, that the shedding region have no effect on the pressure and Reynoldsnumber in the interesting region ?
It's look like a steady state solution from a transient run, is this right ?

I'm just surprised about the fast convergence ?!
Attached Images
File Type: jpg residuals.jpg (29.2 KB, 4 views)
File Type: jpg imbalances.jpg (24.6 KB, 6 views)
File Type: jpg pressure_monitor_points.jpg (38.6 KB, 5 views)
File Type: jpg reynolds.jpg (20.5 KB, 6 views)
biro is offline   Reply With Quote

Old   November 3, 2014, 07:48
Default
  #26
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
Sure looks steady state to me. So it looks like you needed a transient run to get to a steady state solution. This happens sometimes.
ghorrocks is offline   Reply With Quote

Old   November 3, 2014, 08:42
Default
  #27
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
Now I'm happy, after a long time I have finally a solution.


Thank you for you endurance, ghorrocks. It was a great help!
biro is offline   Reply With Quote

Old   November 14, 2014, 13:09
Default
  #28
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
Once again I have a question about how to interpret my transient run.

My tutor said, the fluctuations of the MassFlow (it's MassFlow at inlet) from the steady state run (see picture) have also to occur in the transient run. I don't understood this, because I have no transient boundary.

Can someone explain me, what I really (have to) see when I look at the transient run?

And when I look on the timestep:
The adaptive timestep method decrease the timestep from initial timestep (4e-4) to the minimum timestep (1e-10).
I thought the adaptive timestep method finds the rigth timestep to resolve transient flow changes ?

Now I'm a little bit confused.
Attached Images
File Type: jpg massFlow_steady_state.jpg (23.2 KB, 2 views)
File Type: jpg imbalances_steady_state.jpg (39.4 KB, 2 views)
File Type: jpg rms_steady_state.jpg (38.3 KB, 2 views)
File Type: jpg massFlow_transient.jpg (19.0 KB, 1 views)
File Type: jpg imbalances_transient.jpg (25.1 KB, 1 views)
biro is offline   Reply With Quote

Old   November 14, 2014, 13:09
Default
  #29
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
Here are also the RMS values for the transient run.
Attached Images
File Type: jpg rms_transient.jpg (36.0 KB, 3 views)
biro is offline   Reply With Quote

Old   November 15, 2014, 00:47
Default
  #30
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
What settings did you use for the adaptive time stepping?
ghorrocks is offline   Reply With Quote

Old   November 15, 2014, 20:36
Default
  #31
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
I made screenshots of the settings I have used
Attached Images
File Type: jpg settings_adaptive_1.JPG (58.6 KB, 7 views)
File Type: jpg settings_adaptive_2.jpg (38.7 KB, 8 views)
File Type: jpg settings_solver_control.jpg (46.5 KB, 6 views)
biro is offline   Reply With Quote

Old   November 16, 2014, 05:05
Default
  #32
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
Of course your simulation went bezerk - you have 1e-12 as a residuals convergence criterion. You should replace this with something more sensible, 1e-4 is a good starting point for a transient simulation.
ghorrocks is offline   Reply With Quote

Old   November 16, 2014, 06:16
Default
  #33
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
I choose the value, because I want to stop the run by my self, equal to steady-state run.

But I will start my simulation with a value of 1e-4 once more.
biro is offline   Reply With Quote

Old   November 16, 2014, 06:22
Default
  #34
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
The whole idea of a transient run is to advance it through time. When you choose 1e-12 residual tolerance that means the adaptive time stepping will just reduce the time steps to nothing and you do not advance in time - which defeats the purpose of the transient simulation you are trying to do.
ghorrocks is offline   Reply With Quote

Old   November 20, 2014, 05:20
Default
  #35
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
The timestep size (6.6127e-5) is unchanged since 250 timesteps. Can I assume, that a good timestep size is found ?

Is it recommend to start another run with a fix timestep size found by the adaptive Method ? Or should I change the residual convergence criteria ?


You will see, that all flow values begin to oscillate. I want to research the frequencies and the different values at different timesteps.
Attached Images
File Type: png rms.png (18.7 KB, 2 views)
File Type: png pressure_monitor_points.png (52.3 KB, 2 views)
File Type: png Reynolds_inlet.png (18.9 KB, 3 views)
biro is offline   Reply With Quote

Old   November 25, 2014, 10:21
Default
  #36
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
Does no one has an answer ? My simulation is advanced very well.

What can I do for further investigation ? Should I run with fix timestep size ? Or shoud I increase residual tolerance ?
Attached Images
File Type: png rms.png (17.3 KB, 1 views)
File Type: png pressure_monitor_points.png (49.4 KB, 1 views)
File Type: png Reynolds_inlet.png (17.0 KB, 2 views)
biro is offline   Reply With Quote

Old   November 25, 2014, 18:31
Default
  #37
Super Moderator
 
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,844
Rep Power: 144
ghorrocks is just really niceghorrocks is just really niceghorrocks is just really niceghorrocks is just really nice
Your results show there is clearly a strong transient effect happening.

But what happens next is entirely up to you. Do you want to investigate the periodic cycle? Or the startup transient? Or are you just interested in the overall averaged behaviour? Each of these will require a different approach.
ghorrocks is offline   Reply With Quote

Old   November 26, 2014, 06:02
Default
  #38
Member
 
Robert Bischof
Join Date: Aug 2014
Posts: 56
Rep Power: 12
biro is on a distinguished road
Initially I'm interested in the behaviour of the periodic cycle (frequences and maybe pressure allocation on different times) and on the overall averaged behaviour.
So I will execute the items one by another, but I don't really know how to start for both.

For a start I would disregard the transient startup.
biro is offline   Reply With Quote

Reply


Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
CFX vs FLUENT for particle tracking simulation iman ANSYS 2 August 31, 2012 21:34
nucleate boiling simulation in CFX Anil CFX 3 August 25, 2010 15:18
CFX steady simulation gharek CFX 1 April 7, 2010 19:41
2D simulation - ICEM meshing for CFX question Ben Makhal CFX 5 April 11, 2007 09:44
Simulation of turbine cascade in CFX. Jonas Pedro Caumo CFX 0 December 9, 2006 14:54


All times are GMT -4. The time now is 12:25.