Solution settings for soot model (method of moments)

Hi,

I have set up a reacting flow 2D-model (asymmetric-swirl) that includes of GRI-3 mechanism, turbulence, radiation and soot.

I tried to solve it step by step, from cold flow to addition of the modules one by one. Now in the complete model, after around 500k iterations, the divergence on soot model is occurring. I tried to reduce the URF for soot model, but it is not helping and divergence happens soon or late!

The message I get is as below:

Stabilizing soot-mom-2-moment to enhance linear solver robustness.
        Stabilizing soot-mom-2-moment using GMRES to enhance linear solver robustness.

Divergence detected in AMG solver: soot-mom-2-moment

I also tried to set the solver settings for soot, from flexible to V-cyle with BGSTAB, but it did not help.

My first question: is 'soot-mom-2-moment' about 'soot moment 2' itself? or is it about interaction between soot and momentum?

the plot of residuals at this step is as follows, where the soot-mom-0 shows fluctuation, but the divergence message is about soot-mom-2:

Here are the URFs I've set:

pressure, density: 0.3

body forces, turbulence factors, momentum: 0.7

energy, P1: 0.9

soot: 0.6 (I tried 0.5, 0.4, they also diverge after some more iterations)

What do you think about the origin of this divergence? and how you think I can make the solution stable?

Many thanks in advance for your ideas,

Zari

Comments

  • ZariZari Member
    edited September 2019

    I  disabled the soot model and continued the calculation (soot equations unchecked in Controls/equiations),

    and it still diverges with the same message, soot-mom-2-moment

  • ZariZari Member
    edited September 2019

    Something is wrong with the system behind this community.

    I posted my question around 10 days ago, 0 views until now, and it was only me who added one more thing to it,

    then I received an automatic email that:

    Your topic Solution settings for soot model (method of moments) on ANSYS Student Community has 2 replies but has still not been marked as solved. Please check again to see if any answers are appropriate and mark them as the solution if they are.

    !

  • RobRob UKForum Coordinator
    edited September 2019

    Not really, but if only staff answer questions you need to be patient if we're away: the more questions the community can answer the more time we have for the difficult ones. Staff also cannot mark/like an answer so unless the OP does this the thread will simply drop off the page. 

    Anyway, if you're using the method of moments approach, yes the error is related to the moments. Reading through the theory, it's likely instabilities are caused by too rapid a change in soot concentration or size.  If you revert to one of the other models how does the soot concentration look?  

  • ZariZari Member
    edited October 2019

    Thank you so much for the reply,

    I know that there are many many questions from all users around the world. But at the same time It would be very interesting if the atmosphere of this community could be more dynamic.

    Regarding my model, I changed the soot model from MOM to Moss-Brookes to see how the results for soot looks like. I will write back when I have some data.

    Meanwhile, what I think about the MOM is also about the parameters that might make that instability. since as I checked the results for soot volume fraction and mass fraction, they are in the scale of 1e+12!!! so, something should not be correct here. Therefore I summarize the settings of the soot model:

    soot model: method of moments with 3 moments,

    mechanism: Built-In HACA

    Nucleation: using precursor C2H2,

    site density: 2.3e+15 (1/cm2)

    OH model: instantaneous

    soot radiation interaction is enabled.

    the default parameters and coefficients are kept:

    Process parameters, mean density of soot particle: 2000 kg/m3

    Normalization parameter for moments: 1e+9 kg/m3

    Aggregation, critical diameter: 1e-8 m,

    Fractal diameter: 1.8

    I am suspicious about the value for the normalization parameter for moments. To see its effect on the calculation stability, I tried other values like (1e+7, and 1e+11) but it did not affect the results or stability of the solver.

    Thanks for your time and support,

    Zari/

  • RobRob UKForum Coordinator
    edited September 2019

    Thanks for offering to help: I look forward to reading your posts answering questions. 

    Try dropping the site density. I don't use the model but suspect if you have too high a value to create soot it'll give silly answers. 

Sign In or Register to comment.