FDS is a low Mach number model, so avoid exit speeds that approach the
speed of sound. Most often, we just use a larger orifice and lower
velocity. Also, in default mode, FDS does not account for high strain
extinction, like blow-off. Finite-rate chemistry is possible, but
often difficult when the resolution of the grid is not adequate.
Best advice is to try something simple and work your way up. The issue
is not necessarily whether FDS can "do" a particular calculation, but
rather how well. You can certainly shoot methane out of a hole and
make it burn, but whether this is a good model of the real thing
depends on a number of parameters. Jason Floyd did some work on
hydrogen jets with FDS for the NFPA Research Foundation. You might
want to see if that report is available.