8
$\begingroup$

I am not entirely sure if this is better suited to Chemistry.SE or Matter modelling, but hey, you guys are nicer.

I am fascinated by just how complicated combustion processes are.

I was watching an old SpaceX video on how they model the combustion processes in their Raptor engines. (Relevant link with timestamp: https://www.youtube.com/watch?v=txk-VO1hzBY&t=39m40s, a really excellent video btw, which I recommend people watch in full.)

They give the example of hydrogen burning with oxygen, which requires 23 separate chemical equations to model, and that is with the reduced mechanism.

The myriad of intermediate chemical reactions involved in combusting hydrogen with oxygen

(Incidentally, I would like to know what column 'b' refers to in the table above.)

The situation with methane combustion is even worse. 53 species and 325 reactions are involved.

If one were to attempt to model simplified versions of these processes, how would they go about it?

Is there an a-priori way of telling which individual reactions contribute little to the overall combustion, just by the identity of the species involved and the other entries in the table? Intuitively, one expects the interactions of two different rare high-energy intermediates to not contribute greatly to the overall combustion, but what if it does?

While highly nonlinear, I don't believe these equations are chaotic, at least in a situation with perfect mixing. Even so, I suspect changes in the proportion of trace intermediates could very easily have large effects on subsequent reactions.

Does it come down to running the full simulation, then successively removing reactions until the accuracy drops below an acceptable threshold? Or are there tricks one can use to simplify these equations? Is it just a matter of doing the full simula

$\endgroup$
6
  • $\begingroup$ I deleted my comment after I watched the video, since it is about kinetic modeling. $\endgroup$
    – B. Kelly
    Aug 15 at 2:27
  • 1
    $\begingroup$ However, for chemical reaction equilibria, you choose the set of reactions that works best for your numerical solver. Some folks change the reaction sets on the fly in their solver. $\endgroup$
    – B. Kelly
    Aug 15 at 2:28
  • $\begingroup$ I don't know much about this so I hesitate to answer, but one approach is to use a master equation: en.wikipedia.org/wiki/Master_equation . In that approach, you have to have some idea of the rate of moving between any two of the states described by the reactions you've listed. This rate is then related to a probability that enters a transition matrix, and the entire system is modelled via a Markov process. This allows you to find out how various initial conditions will behave over time and what possible equilibria can be achieved. $\endgroup$
    – jheindel
    Aug 16 at 21:32
  • $\begingroup$ @ingolifs would you be interested in my putting how to determine a set of reactions for chemical reaction equilibria, how there are infinite possibilities, and how people more in the business than myself choose their sets of reactions? It wouldn't answer your question directly, but I would include a monologue about how SpaceX is probably doing it the overly hard way and how that doesn't surprise me :) $\endgroup$
    – B. Kelly
    Aug 19 at 13:54
  • 1
    $\begingroup$ +1 (I gave it earlier, but didn't get a chance to comment at that time). Welcome to our new community and thank you so much for contributing your question here! We hope to see much more of you in the future !!! Here's a related (though not the same) question that I answered on this site: mattermodeling.stackexchange.com/a/6584/5. I think your question could be answered by this community if we wait long enough, but a bounty would probably get people more motivated to put the effort in! $\endgroup$ Aug 21 at 18:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.