2D Rotating Detonation Engine Periodic Boundaries

whitet86 · June 24, 2015, 17:26

Hello everyone,

I have been working on replicating and modifying some recent papers on rotating detonation engines by Escobar et al and Hayashi et al in Fluent 15. To do so, I basically unwrapped the annular combustion chamber into a 2d rectangular, axisymmetric domain, and assumed premixed flow into the inlet and and out of the outlet with periodic conditions on the remaining boundaries to simulate a repeated circular motion. I used a mapped face grid since it was a simple rectangular boundary, the skewness was e-10 and grid quality at 1 or nearly. It is also using the density based solver because of the detonations.

My problem is this fails to even initialize once I apply the periodic boundary conditions. I receive an error on the first iteration of Divergence error ads-0 and error #f.

Methods I have attempted to create these boundaries:
I have attempted to do these on the X-Y plane as horizontally and vertically, and on the X-Z plane because I have read that FLuent requires certain conditions to be modeled in certain directions.

In ICEM I have created a cylindrical coordinate system and attempted to make them cyclic. Once it loaded by returned the same error, and other times this won't even grid properly and returns an error say elements were moved farther than the tolerance. It will not allow me to use periodic boundaries on a 2d model for some reason as it requires face selection for both low and high boundaries.

In Fluent I have attempted the set the walls as interfaces, and then matched these interfaces as periodic boundaries. It returned the above error.

In Fluent I have used the TUI commands to make it periodic. It returned the same error. I have tried to set rotational directions in the cell zone conditions but that didn't work.

If I can't get these boundary conditions to even initialize, I cannot move forward into patching in the combustion. Has anyelse had any experience with these boundary conditions for 2D or the ads-0 error?

Thank you all,
Tim

sescobar · June 24, 2015, 22:13

Hi Tim,
Unwraping the RDC annular chamver leads to a 2D problem with periodic boundaries. This is different from axisymmetric. Just make sure than none of your boundaries are defined as axis.
The periodic boundary condirion can be defined using ICEM or Fluent. I normally defined them as walls in ICEM and change them to periodic in Fluent. You need to make sure that the periodic boundaries have corresponding nodes.
Try this and let me know if this works.

Also if you want to, send me your email address and we can exchange contact information.

Best regards,

Sergio Escobar

whitet86 · June 25, 2015, 11:04

Wow thank you Dr. Escobar. You were possibly the last person I expected to respond lol. I will try your suggestion immediately.

June 24, 2015, 17:26	2D Rotating Detonation Engine Periodic Boundaries	#1
whitet86 New Member Join Date: Jun 2015 Posts: 7 Rep Power: 10	Hello everyone, I have been working on replicating and modifying some recent papers on rotating detonation engines by Escobar et al and Hayashi et al in Fluent 15. To do so, I basically unwrapped the annular combustion chamber into a 2d rectangular, axisymmetric domain, and assumed premixed flow into the inlet and and out of the outlet with periodic conditions on the remaining boundaries to simulate a repeated circular motion. I used a mapped face grid since it was a simple rectangular boundary, the skewness was e-10 and grid quality at 1 or nearly. It is also using the density based solver because of the detonations. My problem is this fails to even initialize once I apply the periodic boundary conditions. I receive an error on the first iteration of Divergence error ads-0 and error #f. Methods I have attempted to create these boundaries: I have attempted to do these on the X-Y plane as horizontally and vertically, and on the X-Z plane because I have read that FLuent requires certain conditions to be modeled in certain directions. In ICEM I have created a cylindrical coordinate system and attempted to make them cyclic. Once it loaded by returned the same error, and other times this won't even grid properly and returns an error say elements were moved farther than the tolerance. It will not allow me to use periodic boundaries on a 2d model for some reason as it requires face selection for both low and high boundaries. In Fluent I have attempted the set the walls as interfaces, and then matched these interfaces as periodic boundaries. It returned the above error. In Fluent I have used the TUI commands to make it periodic. It returned the same error. I have tried to set rotational directions in the cell zone conditions but that didn't work. If I can't get these boundary conditions to even initialize, I cannot move forward into patching in the combustion. Has anyelse had any experience with these boundary conditions for 2D or the ads-0 error? Thank you all, Tim miladz and Bonks like this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
problem with cyclicAMI periodic boundaries inside MRF zone	windmachines	OpenFOAM Running, Solving & CFD	16	September 12, 2016 21:57
No flow through periodic (cyclic) boundaries in impeller with foam-extend-3.1	anttiad9000	OpenFOAM Running, Solving & CFD	3	March 2, 2016 19:37
Pulse Detonation Rocket Engine	aero_peace	FLUENT	2	November 28, 2014 12:01
[ICEM] Help with periodic boundaries	Mitpostdoc	ANSYS Meshing & Geometry	3	March 24, 2011 16:57
[ICEM] Periodic Boundaries for Hexa/Tetra mesh	sebastian	ANSYS Meshing & Geometry	0	December 23, 2009 05:57

June 24, 2015, 22:13		#2
sescobar New Member Join Date: Jun 2011 Posts: 22 Rep Power: 14	Hi Tim, Unwraping the RDC annular chamver leads to a 2D problem with periodic boundaries. This is different from axisymmetric. Just make sure than none of your boundaries are defined as axis. The periodic boundary condirion can be defined using ICEM or Fluent. I normally defined them as walls in ICEM and change them to periodic in Fluent. You need to make sure that the periodic boundaries have corresponding nodes. Try this and let me know if this works. Also if you want to, send me your email address and we can exchange contact information. Best regards, Sergio Escobar

June 25, 2015, 11:04		#3
whitet86 New Member Join Date: Jun 2015 Posts: 7 Rep Power: 10	Wow thank you Dr. Escobar. You were possibly the last person I expected to respond lol. I will try your suggestion immediately.