I have bumped into quite an interesting problem. The problem is to do combustion analysis in a closed chamber, I mean imagine an internal combustion engine problem, the fuel and air are mixed and left combusting in a closed chamber. The problem here is in this case there is neither inlet nor outlet. Does anybody has experience / clue how to go about this kind of problem. Comments are highly appreciated.

This is a highly comlex problem as you need to consider phenomena like ignition and extinction. Of the combustion models present in Fluent, only EDC and PDF transport may be used for such a problem. You need to use detailed chemistry and it'll cost you a LOT in terms of CPU resources.

Your experiences with use of Fluent for this analysis would be highly interesting though. Good luck!

actually i have a good lot of experince with combustion models (i mean with EDC and pdf transport), but my prob here is how to set up a case and run with no inlet and outlet, but yet assuming that we have a combustion mixture in the domain.

Hi, maybe you don't need Fluent to solve this problem.

If the chamber is closed (no inlets and no outlets), and the reactants are mixed, you can consider all the properties and composition uniform in the volume, and treat the system as a constant-volume batch reactor.

You just need to follow the dynamics (i.e. the variation with time) of the quantities you need to know, starting from the initial conditions.

If you are interested in more information, such as pressure waves or evolution of temperature profiles as a consequence of a spark, things become more complex and you'll need to solve PDEs; however, I am not expert in this field, but I think that you could still solve the problem without Fluent.

Anyway, It could be interesting to know if it is possible to solve your problem by Fluent simply setting properly the boundary conditions.

Regards,

Ale

actually i also feel that i need to go for other options than fluent, luckly i have been working on a solver code for last few months, and hope to get the solver running of unstruct-grids (ie to import the grid from gambit) by the end of this month. then i shall be able to find out whether this will be possible. but still i need to get some idea how people (at least who has some experience in this field) feel about approaching this problem as this could be great help in getting solution. or at least if some can give some pointer to read about such probs.

You don't need inlet and outlet for this kind of analysis, as it is inherently a time-dependent problem. You only need to start things moving, i.e. to supply the "spark" and this is generally done by patching a very small region with a high temperature.

There are not so many publications on IC engines in the open literature that would describe this in detail, but I believe you will still be able to find sufficient information without much trouble.

humm, i will give it a try and lets see what happens.