The convergence of the solution is not observed (HiSA)
Moderator: oliveroxtoby
Forum rules
and Helpful information for the FEM forum
and Helpful information for the FEM forum
The convergence of the solution is not observed (HiSA)
Hello. My task is to get a view of the flow in the nozzle at a pressure of 3 MPa and temperature of 400 K in the chamber and 0.1 MPa, 297 K outside. There is also a second inlet above the nozzle cutoff with 0.1 MPa pressure and 297 K temperature (injection). I use HiSA with OpenFOAM v2112 at high Courant numbers for fast convergence (a necessary condition). However, I observe poor convergence in terms of both residuals and resulting fields (very strange behavior behind the shockwave) .
This problem is known to have a stationary solution, since the BCs and type of geometry were taken from a problem solved in Ansys Fluent, where the solution had a particular form. (Outlet area was increased).
There is case of the problem: https://drive.google.com/file/d/1oMRoYo ... sp=sharing (updated). Also this case includes .avi and residuals file to control them with gnuplot.
UPD:
I have attached an example with a given grid and fields for the solution, as well as the required flow shape (Ansys).
The animation of solution may be recieved there: https://drive.google.com/file/d/1jnfc8v ... sp=sharing.
I think the problems are related to the boundary conditions. It has taken me a long time to figure out what the problem is, but I still haven't figured it out. Could you please help?
Mach distribution and streamlines of velocity are respresented in attachments (which should be received): Also there are residuals and other for understanding:
This problem is known to have a stationary solution, since the BCs and type of geometry were taken from a problem solved in Ansys Fluent, where the solution had a particular form. (Outlet area was increased).
There is case of the problem: https://drive.google.com/file/d/1oMRoYo ... sp=sharing (updated). Also this case includes .avi and residuals file to control them with gnuplot.
UPD:
I have attached an example with a given grid and fields for the solution, as well as the required flow shape (Ansys).
The animation of solution may be recieved there: https://drive.google.com/file/d/1jnfc8v ... sp=sharing.
I think the problems are related to the boundary conditions. It has taken me a long time to figure out what the problem is, but I still haven't figured it out. Could you please help?
Mach distribution and streamlines of velocity are respresented in attachments (which should be received): Also there are residuals and other for understanding:
Last edited by Reptider on Sun Dec 11, 2022 4:56 pm, edited 1 time in total.
-
- Veteran
- Posts: 3129
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: The convergence of the solution is not observed (HiSA)
This is what I get:
I put the inlet away from the nozzle, so that v/p/T-fields are better calculated
at the nozzle area. I dont know how to interpret the residuals and the solver message.
Is that a good sign?
Meshing is under 2 min, you must adjust the number of cpu-cores in the CFD_solver Data-tab.
Inlet speed is 30 m/s, no turbulence model.
I put the inlet away from the nozzle, so that v/p/T-fields are better calculated
at the nozzle area. I dont know how to interpret the residuals and the solver message.
Is that a good sign?
Meshing is under 2 min, you must adjust the number of cpu-cores in the CFD_solver Data-tab.
Inlet speed is 30 m/s, no turbulence model.
-
- Veteran
- Posts: 3129
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: The convergence of the solution is not observed (HiSA)
The simulation runs much longer (1200 iterations now, 600 before), when the outlet-domain is extended.
At a certain point Hisa suddenly stopps. This is the last iteration you see here.
Maybe the mesh is still too coarse for a good resolution of the shockwaves.
But I find it quite amazing, that Hisa can compute the "diamonds".
Cheers!
At a certain point Hisa suddenly stopps. This is the last iteration you see here.
Maybe the mesh is still too coarse for a good resolution of the shockwaves.
But I find it quite amazing, that Hisa can compute the "diamonds".
Cheers!
- oliveroxtoby
- Posts: 812
- Joined: Fri Dec 23, 2016 9:43 am
- Location: South Africa
Re: The convergence of the solution is not observed (HiSA)
This was my mistake. The open pressure boundary was not selecting the appropriate all-Mach boundary condition for HiSA. I have fixed this and tested it on the file from @thschrader which seems to work now. Please update your workbench and try again.Reptider wrote: ↑Fri Dec 09, 2022 8:58 am Hello. My task is to get a view of the flow in the nozzle at a pressure of 3 MPa and temperature of 400 K in the chamber and 0.1 MPa, 297 K outside. There is also a second inlet above the nozzle cutoff with 0.1 MPa pressure and 297 K temperature (injection). I use HiSA with OpenFOAM v2112 at high Courant numbers for fast convergence (a necessary condition). However, I observe poor convergence in terms of both residuals and resulting fields (very strange behavior behind the shockwave) .
-
- Veteran
- Posts: 3129
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: The convergence of the solution is not observed (HiSA)
Here are the diamonds...
https://en.wikipedia.org/wiki/Shock_diamond
https://en.wikipedia.org/wiki/Shock_diamond
-
- Posts: 4
- Joined: Tue Aug 01, 2023 5:48 am
Re: The convergence of the solution is not observed (HiSA)
Hı dear schrader which boundary condition do you use when you simulate con di nozzle? I have a problem when simulate con di nozzle. Did you use pressure outlet condition end of the domain? Also do we have to enter velocity parameter at pressure inlet ? I want to enter only pressure value at pressure inlet? Could you share your boundary condition ?
Best regards
Best regards
Last edited by ahmetcanmaden on Tue Aug 01, 2023 6:05 am, edited 1 time in total.
-
- Veteran
- Posts: 3129
- Joined: Sat May 20, 2017 12:06 pm
- Location: Germany
Re: The convergence of the solution is not observed (HiSA)
Here are 2 live-cases:ahmetcanmaden wrote: ↑Tue Aug 01, 2023 6:04 am Hı dear schrader which boundary condition do you use when you simulate con di nozzle? I have a problem when simulate con di nozzle. Did you use pressure outlet condition end of the domain? Also do we have to enter velocity parameter at pressure inlet ? I want to enter only pressure value at pressure inlet? Could you share your boundary condition ?
Best regards
viewtopic.php?t=79066
viewtopic.php?p=541358&hilit=blueorigin#p541358
You can find the rocket example in your cfdof installation folder, the projectile.Run the macro.
and
-
- Posts: 4
- Joined: Tue Aug 01, 2023 5:48 am
Re: The convergence of the solution is not observed (HiSA)
Thank you for replying dear schrader but i specificaly need nozzle inlet boundary condition for your convergent divergent nozzle analysis. İn fluent we dont have to enter velocity parameter at nozzle inlet we only enter pressure and temperature parameter. But hisa wants to enter velocity parameter at nozzle inlet . I winder how did you enter the boundary condition at nozzle inlet? Best regards
- oliveroxtoby
- Posts: 812
- Joined: Fri Dec 23, 2016 9:43 am
- Location: South Africa
Re: The convergence of the solution is not observed (HiSA)
The velocity is only used for any inlet faces that exceed Mach 1. So most likely it can be ignored.ahmetcanmaden wrote: ↑Thu Aug 03, 2023 6:41 am Thank you for replying dear schrader but i specificaly need nozzle inlet boundary condition for your convergent divergent nozzle analysis. İn fluent we dont have to enter velocity parameter at nozzle inlet we only enter pressure and temperature parameter. But hisa wants to enter velocity parameter at nozzle inlet . I winder how did you enter the boundary condition at nozzle inlet? Best regards
-
- Posts: 4
- Joined: Tue Aug 01, 2023 5:48 am
Re: The convergence of the solution is not observed (HiSA)
Thank you for replying dear Oliver. I guess I can't express myself correctly. I want to ask which boundary did you use at nozzle upstream inlet. İn hisa there is only characteristicpressureinletoutletvelocity boundary condition and this boundary condition requires velocity input but i dont know the nozzle upstream inlet velocity. Which velocity parameter should i enter the nozzle inlet boundary? Best regards