Modelling reversible chemical reaction.

kraviec · August 20, 2010, 09:19

Hello,

I’m trying to model a liquid-liquid reversible reactions. What is more that is axissymetric problem with swirl. I have some problems, because I’m relatively new in Ansys Fluent.
To start, my reactions are as follows:

1A + 1B <-> 1C + 1D
1C + 1B <-> 1E + 1D

I have all the data for reaction, k constant and activation energy. I’ve calculated pre-exponential factor based on Arrhenius equation. Also I got all required data in materials panel. I know that constant k for reversible reaction is calculated from standard state entropy and enthalpy.
First question is what is rate exponent and how to calculate it? Chemical kinetics is not my best side, so I don’t know, if I calculated it properly.

Still it won’t work properly. I use unsteady solver, laminar flow (there is no need to use turbulence model - it is based on known Reynolds number), 2ddp. In species transport, there is used volumetric type of reaction (laminar finite rate).

Another question is about species transport. At the outflow boundary, there is backflow of last (E) compound (I know that the last one is complementary compund). The same situation was, when I was using pressure-outlet type of boundary. Any ideas?

Last but not least question, is about mesh and time step size. Do these problems can result from too small amounts of mesh elements or too large time step, whether to few iteration per time step?

Thanks in advance for all kinds of ideas or suggestions.

Allan Walsh · August 20, 2010, 12:08

Often the rate exponents for the reactants will be the same as the stoichiometric coefficient in the reaction. For example, in the reaction
Ca + 0.5 O2 -> CaO, then
rate = k [Ca] [O2]^0.5.
If you follow the crude computer notation for the equation, then the rate exponent for oxygen is 0.5.
In practice it can be much more difficult. You should research your specific equations to see if there are suggested literature values for a mechanism with rate exponents. Also, the Fluent (Ansys) materials database has quite a few built-in reaction mechanisms and you may get lucky with yours.
In my experience, it is easy to make mistakes on the units for the pre-exponential factors and I recommend doing a hand calculation as a check of the rate coefficients calculated by Fluent.

kraviec · August 20, 2010, 13:49

The reaction is transeterification and accurate data is hard to find, because it happens in spinning disk reactor (we have only data from batch reactor) and that is a new problem for us.

I've done hand calculation for pre-exponential factor. Based on Fluent help, I found that reaction k constant units are m^3/(kmol*s) so pre-exponential factor should have consitnt units. Am I thinking right?

Also may I ask if you could tell me how to change built-in reaction mechanism, because I think I'm using a default one?

August 20, 2010, 09:19	Modelling reversible chemical reaction.	#1
kraviec New Member Marek Join Date: Mar 2010 Location: Poland Posts: 2 Rep Power: 0	Hello, I’m trying to model a liquid-liquid reversible reactions. What is more that is axissymetric problem with swirl. I have some problems, because I’m relatively new in Ansys Fluent. To start, my reactions are as follows: 1A + 1B <-> 1C + 1D 1C + 1B <-> 1E + 1D I have all the data for reaction, k constant and activation energy. I’ve calculated pre-exponential factor based on Arrhenius equation. Also I got all required data in materials panel. I know that constant k for reversible reaction is calculated from standard state entropy and enthalpy. First question is what is rate exponent and how to calculate it? Chemical kinetics is not my best side, so I don’t know, if I calculated it properly. Still it won’t work properly. I use unsteady solver, laminar flow (there is no need to use turbulence model - it is based on known Reynolds number), 2ddp. In species transport, there is used volumetric type of reaction (laminar finite rate). Another question is about species transport. At the outflow boundary, there is backflow of last (E) compound (I know that the last one is complementary compund). The same situation was, when I was using pressure-outlet type of boundary. Any ideas? Last but not least question, is about mesh and time step size. Do these problems can result from too small amounts of mesh elements or too large time step, whether to few iteration per time step? Thanks in advance for all kinds of ideas or suggestions.

August 20, 2010, 12:08		#2
Allan Walsh Member Allan Walsh Join Date: Mar 2009 Posts: 56 Rep Power: 17	Often the rate exponents for the reactants will be the same as the stoichiometric coefficient in the reaction. For example, in the reaction Ca + 0.5 O2 -> CaO, then rate = k [Ca] [O2]^0.5. If you follow the crude computer notation for the equation, then the rate exponent for oxygen is 0.5. In practice it can be much more difficult. You should research your specific equations to see if there are suggested literature values for a mechanism with rate exponents. Also, the Fluent (Ansys) materials database has quite a few built-in reaction mechanisms and you may get lucky with yours. In my experience, it is easy to make mistakes on the units for the pre-exponential factors and I recommend doing a hand calculation as a check of the rate coefficients calculated by Fluent. soheil_r7 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFX5.7 - chemical reaction problem...	H.L.	CFX	4	March 25, 2009 10:33
Modelling batch chemical reaction	Frans D	FLUENT	0	October 30, 2007 22:59
Specifying a chemical reaction	hagupta	CFX	3	March 22, 2006 01:46
chemical reaction - decompostition	La S. Hyuck	CFX	1	May 23, 2001 00:07
chemical reaction - howto	Dan Sorensen	Siemens	1	February 21, 2000 09:20

August 20, 2010, 13:49		#3
kraviec New Member Marek Join Date: Mar 2010 Location: Poland Posts: 2 Rep Power: 0	The reaction is transeterification and accurate data is hard to find, because it happens in spinning disk reactor (we have only data from batch reactor) and that is a new problem for us. I've done hand calculation for pre-exponential factor. Based on Fluent help, I found that reaction k constant units are m^3/(kmol*s) so pre-exponential factor should have consitnt units. Am I thinking right? Also may I ask if you could tell me how to change built-in reaction mechanism, because I think I'm using a default one?