Learning Forum

[Biborka Boga]

Hi! Before describing the problem, my question would be: can you please offer me a feedback regarding to the correctness of the following proposed approach? Thank you very much in advance!

I am involved in the modeling of L-S mass-transfer of dissolved O2 (from the liquid phase to the catalyst, which in my case is defined as a wall). Based on the previous discussions and based on literature study I have concluded that there are two options:

(1) introducing the adsorption as a wall reaction (in the last few months I have been insisting on this approach inclusively on site and bulk species, finding appropriate activation energies, preexponential factors, however after a lot of tryings I have given up this approach)

(2) Defining the catalyst as a sink term of certain species (via user defined scalars)

Although I have a multiphase problem, for simplicity now I have modelled just the liquid film. Let’s consider the mixture of H2O+O2 (with an initial molar fraction of dissolved O2: 1.62e-10). At the upper wall BC I have defined a boundary O2 flux in order to assure the saturation of H2O with O2 (please find below) along the reactor length.

 

Aim: defining the catalyst layer a sink term for dissolved O2 (however the aim is to extend the model to other species)

The conservation of O2 can be written as depicted in (1):

where  S is the source/sink term.

Although I am aware of the fact that the following approach is oversimplied, at first I would consider:

Do you think that this approach will work?

Thank you very much in advance your feedback regarding to the correctness of the proposed procedure.

 

 

[Judy Cooper]

Hi Biborka:

 

Your approach may work with a small enough timestep.

Sources that are a function of the solved variable tend to benefit from linearization, however.  This topic is discussed in the help documentation (Fluent customization manual) here:

https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v221/en/flu_udf/flu_udf_sec_define_source.html

 

So you could implement the UDF as follows:

 #include "udf.h"
 
 #define c2 -10000.0
 
 DEFINE_SOURCE(o2_source,c,t,dS,eqn)
 {
    real k,source;
    source = c2*k*C_YI(c,t,i);
    dS[eqn] = c2*k;
    return source;
 } 

 
C_YI is the O2 species mass fraction (i is index from 0-1 given by species order).
c2 is a multiplier that strengthens the effect of the linearization.
dS[eqn] is the source coefficient (usually proportional to the derivative of the source term w.r.t. the solved variable).
This must always be negative.

When the source is applied to the equations, set it as the mass source and the o2 species source to indicate that all the mass the source pulls will be pulled from the O2 species equation.

 

 

[Biborka Boga]

Thank you very much! I am very grateful the answer provided! :)