[sooraj546]

Sir,
I am trying to simulate a problem including three source terms one including temperature and others including 2 different species.

Problem definition:

A combustion chamber with air and fuel inlet and a porous sink on its bottom completes the geometry. Combustion of 8 species happens in the combustion chamber(ch4,co,h2,h20,c2h6,n2,o2,co2)in the presence of air and a temperature and species concentration is developed inside the combustion chamber and sink. After running this in steady state for 5000 iterations. I want to simulate the reaction inside the porous media for 20 min. and find out the temperature and concentration variation of CO and CO2. For this in the bottom porous sink portion i have to add temperature and co and co2 source term. But while doing it i am getting the following error.
# Divergence detected in AMG solver: species-2 -> Increasing relaxation sweeps!
Error: Divergence detected in AMG solver: species-2
Error: Divergence detected in AMG solver: species-2
Error Object: #f
i tried increasing the relaxation sweep but no use same error is dominant . Kindly look at my code and say what the error. My doubts

(1) temperature source term i think is ok as units are matching but my species source term when we do C_YI macro we get the mass fraction nt the concentration in-order to convert it its given use this has someone use udf to add source term? etc. I have used a tech kindly check it

(2) how to overcome this divergance problem

[sooraj546]

its a model concerning EDDY DISSIPATION , DISCRETE ORDINATE, K-E TURBULENCE AND ENERGY EQ AS WELL. kindly help me out. the udf is running but after some time divergence in the species 2 ie o2 is observed i tried increasing the relaxation factor still it exists

[sooraj546]

Kindly someone reply

[CeesH]

Divergence is quite difficult to pinpoint exactly. With all the species, it seems to me you have quite a complex reaction model. It may be something goes wrong in there - did you try a simpler reaction model (or no reaction at all - to ensure divergence is not due to the flow itself)

The chemistry is quite likely stiff since you are dealing with combustion, so are you taking any special measures to deal with that? (and to check - what is your timestep size, and what are the typical timescales of your reactions?)

[sooraj546]

THANKS a lot for that reply sir, On practical basis the reduction happens for 20 min inside the porous zone so i have simulated for 20 min with a time step of 1 sec. any ways i will try it out with a simple model and update you with the results

[CeesH]

It may be that the entire process takes as long, but there are some (very) rapid intermediate steps occurring at second scales - which are then problematic to capture if you use a time resolution of 1 second. You can make an estimate of timescales (1/k in case of a 1st order reaction), and if any very rapid reaction occur, try and eliminate them via quasi-equilibrium assumptions (or use the stiff chemical solver of ANSYS)

[sooraj546]

once again thanks a lot sir for that reply. I will try out all your valuable suggestions and update you with the results