nonequilibrium testcase with 11 species

eli123 · December 28, 2021, 05:03

Hello,

I have completed the case viscwedge_mpp with the version SU2 NEMO 7.2.1 then i made changes to the cfg file to use the air_11 mutation model. After 8 iterations the code stops. I am attaching cfg file along with the screen dump, Any suggestions would be very helpful.

bigfootedrockmidget · December 28, 2021, 07:19

Are you sure that this is the config file that you are using? In the SU2 output it says:

Code:

Maximum number of solver subiterations: 11.

and also:

Code:

Maximum number of iterations reached (ITER = 11) before convergence.

but your config file has:

Code:

% Number of total iterations
ITER= 1000

eli123 · December 28, 2021, 09:04

Correct. I updated the number of iterations to get a more converged solution.

wallym · December 29, 2021, 15:47

Hi eli123,

My guess is that mesh being used is not high-enough quality for an 11-species gas, it is mainly there to run the regression cases. I would suggest refining the mesh, if possible.

Best,
Wally

eli123 · January 6, 2022, 06:07

Hi,

Thank you for your prompt answer.
I tried to refine the mesh at an acceptable level, also decreased the CFL and finally ran to first order but the simulation failed again.
By the way is there any available tutorial using 11 species and Mutation++ ?

eli123 · January 6, 2022, 06:38

...by decreasing the CFL to a value of 2e-3 the computation seems to continue

wallym · January 7, 2022, 13:26

Hi,
That is an awfully low CFL number. Let me try this case on my end to see what may be happening. That behavior isn't right.
EDIT/UPDATE: I found an issue that may be causing the issue and a pull request is up on the github. Hopefully the issue will be resolved soon.

As for tutorials, the NEMO team is working on creating them in the next few months.

eli123 · January 9, 2022, 04:32

Hi,

thank you for the bug update.
The correction is relative to EULER IMPLICIT time scheme but as for now only EULER EXPLICIT time scheme is available with NEMO, right ?

wallym · January 9, 2022, 13:51

Hi,
Both implicit and explicit are available in NEMO, at least in the develop branch of the code. As for the fix, there was an issue where the viscous Jacobians were being used even if the code was using the EXPLICIT formulation.

eli123 · January 10, 2022, 03:45

Hi,

I used the develop version and got the following error message :
"Implicit time scheme is not yet implemented with Mutation++. Use EULER_EXPLICIT"
So, how can I run a simulation with 11 species ?

Any help will be appreciated.

wallym · January 10, 2022, 12:45

Hi,

I apologize, I misspoke. The IMPLICIT solver is only available for SU2's native thermochemical library (SU2_NONEQ). Mutation++ is compatible yet.

eli123 · March 14, 2024, 03:17

Hi Wally,

I downloaded the last version from the 'develop' branch and checked the viscwedge_mpp test case with 11 species.
I cannot converge and after 10 steps the simulation failed.
Can you please help me with this issue.

Regards,
Eli

eli123 · March 19, 2024, 09:10

Hi,

It seems that a CFL of 1e-3 get rid of the convergence problem....

December 28, 2021, 07:19		#2
bigfootedrockmidget Senior Member bigfoot Join Date: Dec 2011 Location: Netherlands Posts: 679 Rep Power: 21	Are you sure that this is the config file that you are using? In the SU2 output it says: Code: Maximum number of solver subiterations: 11. and also: Code: Maximum number of iterations reached (ITER = 11) before convergence. but your config file has: Code: % Number of total iterations ITER= 1000 PENGGEGE777 likes this.

January 7, 2022, 13:26		#7
wallym Senior Member Wally Maier Join Date: Apr 2019 Posts: 123 Rep Power: 7	Hi, That is an awfully low CFL number. Let me try this case on my end to see what may be happening. That behavior isn't right. EDIT/UPDATE: I found an issue that may be causing the issue and a pull request is up on the github. Hopefully the issue will be resolved soon. As for tutorials, the NEMO team is working on creating them in the next few months. Last edited by wallym; January 7, 2022 at 14:34.

January 9, 2022, 13:51		#9
wallym Senior Member Wally Maier Join Date: Apr 2019 Posts: 123 Rep Power: 7	Hi, Both implicit and explicit are available in NEMO, at least in the develop branch of the code. As for the fix, there was an issue where the viscous Jacobians were being used even if the code was using the EXPLICIT formulation. eli123 likes this.

March 19, 2024, 09:10	non equilibrium viscwedge_mpp test case with 11 species	#13
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Hi, It seems that a CFL of 1e-3 get rid of the convergence problem....

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Modelling Biomass Combustion via Species Transport	Racheal	FLUENT	39	January 8, 2022 08:42
UDF for Sorption in Species Transport through Porous Medium	Kosch	Fluent UDF and Scheme Programming	6	October 21, 2020 18:46
Combine Flux/Concentration BC in Electrochemical Species Model	Anodos	STAR-CCM+	0	May 4, 2020 16:03
Site Species	Diger	FLUENT	6	May 3, 2016 06:36
Species order fluent	Raddi	FLUENT	1	March 13, 2009 01:21

December 28, 2021, 09:04		#3
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Correct. I updated the number of iterations to get a more converged solution.

December 29, 2021, 15:47		#4
wallym Senior Member Wally Maier Join Date: Apr 2019 Posts: 123 Rep Power: 7	Hi eli123, My guess is that mesh being used is not high-enough quality for an 11-species gas, it is mainly there to run the regression cases. I would suggest refining the mesh, if possible. Best, Wally

January 6, 2022, 06:07		#5
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Hi, Thank you for your prompt answer. I tried to refine the mesh at an acceptable level, also decreased the CFL and finally ran to first order but the simulation failed again. By the way is there any available tutorial using 11 species and Mutation++ ?

January 6, 2022, 06:38		#6
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	...by decreasing the CFL to a value of 2e-3 the computation seems to continue

January 9, 2022, 04:32		#8
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Hi, thank you for the bug update. The correction is relative to EULER IMPLICIT time scheme but as for now only EULER EXPLICIT time scheme is available with NEMO, right ?

January 10, 2022, 03:45		#10
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Hi, I used the develop version and got the following error message : "Implicit time scheme is not yet implemented with Mutation++. Use EULER_EXPLICIT" So, how can I run a simulation with 11 species ? Any help will be appreciated.

January 10, 2022, 12:45		#11
wallym Senior Member Wally Maier Join Date: Apr 2019 Posts: 123 Rep Power: 7	Hi, I apologize, I misspoke. The IMPLICIT solver is only available for SU2's native thermochemical library (SU2_NONEQ). Mutation++ is compatible yet.

March 14, 2024, 03:17		#12
eli123 New Member anonymous Join Date: Jan 2014 Posts: 9 Rep Power: 12	Hi Wally, I downloaded the last version from the 'develop' branch and checked the viscwedge_mpp test case with 11 species. I cannot converge and after 10 steps the simulation failed. Can you please help me with this issue. Regards, Eli