Machine Learning Integration into PyFluent Solver

[BaileyLadd01]

Hi all,

I'm not going to specifically mention the exact application I am thinking of since I plan on it becoming a research publication in the next year or so.

Has anyone thought of using a machine learning algorithm to predict the affect of chemical reactions on fluid dynamics based on the temperature and/or pressure of a cell in which the reaction occurs? From my understanding (I am just starting out so feel free to correct me), these values are known from the cell calculations. Thus all the machine learning algorithm would do is overwrite the existing chemical reaction to determine the change in fluid properties.

An example to further explain this:
'If you have a basic static mixer which mixes together two fluids, and during this mixing phase there is an observed exothermic chemical reaction. The machine learning algorithm would predict the change in fluid temperature and thus it would be accounted for in the remaining fluid dynamics simulation'

Let me know if this has been looked into, or I have misunderstood something vital about the method that could be applied. Like I said, I won't go into specifics about what exactly I would replace the existing chemical process with.

Cheers,
Vincent