Modeling chemically reacting flow

Zamo · January 18, 2010, 03:02

I am modeling chemically reacting gas flowing through a tube. The reaction is a 5 stage multistage mechanism which i have already incorporated into cfx. Each step has a rate expression. The problem arises when i select the finite rate chemistry option. I then get a overflow error, or a newton iteration error when using a laminar model. I have used a tubulence model (k-e and k-omega hybrid) with no such error, but the solution just returns my initial conditions. Any suggestions?

ghorrocks · January 18, 2010, 05:08

Sounds like the flow is turbulent so a laminar model will never converge (well, not in steady state anyway). Why did you remove the turbulence model? I guess because the model did not move from the initial conditions.

I think you should concentrate on getting the turbulent model to move from the initial conditions.

Zamo · January 18, 2010, 06:06

Im trying to simulate the idealized case before i get ahead of myself but I guess there is no testing the waters. Might as well just jump in. Any suggestions on how to make the solver move from the initial conditions. It is a steady state simulation. After 30 iterations i check the solution and is looks reasonable but the residuals are too high. I then run it from that point on interpolating the results as an initial guess, then it converges to my original initial guess.....what now?

ghorrocks · January 18, 2010, 16:43

It is a common mistake to think that a laminar simulation will be a good start to a more complex turbulent simulation. It does not work that way, you will never get the laminar simulation to converge. The turbulence model is required to damp the turbulence fluctuations to achieve convergence so you cannot run without it.

You have not described what you are doing in much detail so we cannot help your specific issue.

January 18, 2010, 03:02	Modeling chemically reacting flow	#1
Zamo New Member Zamier Join Date: Jan 2010 Posts: 2 Rep Power: 0	I am modeling chemically reacting gas flowing through a tube. The reaction is a 5 stage multistage mechanism which i have already incorporated into cfx. Each step has a rate expression. The problem arises when i select the finite rate chemistry option. I then get a overflow error, or a newton iteration error when using a laminar model. I have used a tubulence model (k-e and k-omega hybrid) with no such error, but the solution just returns my initial conditions. Any suggestions?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Modeling sudden flow in Vacuum chamber	Sandilya Garimella	FLUENT	0	February 8, 2008 15:04
Modeling reduced flow rate in 2D	Tom Smith	FLUENT	2	April 27, 2007 09:04
Modeling Bleed in Supersonic Flow	Mohd Yousuf Ali	FLUENT	0	March 17, 2007 03:51
experts in multiphase flow modeling in fluent.inc?	Asghari	FLUENT	0	March 13, 2007 02:46
Inviscid Drag at subsonic, subcritical Mach #	Axel Rohde	Main CFD Forum	1	November 19, 2001 12:19

January 18, 2010, 05:08		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Sounds like the flow is turbulent so a laminar model will never converge (well, not in steady state anyway). Why did you remove the turbulence model? I guess because the model did not move from the initial conditions. I think you should concentrate on getting the turbulent model to move from the initial conditions.

January 18, 2010, 06:06		#3
Zamo New Member Zamier Join Date: Jan 2010 Posts: 2 Rep Power: 0	Im trying to simulate the idealized case before i get ahead of myself but I guess there is no testing the waters. Might as well just jump in. Any suggestions on how to make the solver move from the initial conditions. It is a steady state simulation. After 30 iterations i check the solution and is looks reasonable but the residuals are too high. I then run it from that point on interpolating the results as an initial guess, then it converges to my original initial guess.....what now?

January 18, 2010, 16:43		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	It is a common mistake to think that a laminar simulation will be a good start to a more complex turbulent simulation. It does not work that way, you will never get the laminar simulation to converge. The turbulence model is required to damp the turbulence fluctuations to achieve convergence so you cannot run without it. You have not described what you are doing in much detail so we cannot help your specific issue.