Divergence during DDT
 

Dear All

I am trying to emulate deflagration to detonation process of ethylene-air mixture in a long tube. The problem is that the simulation runs nicely as long as I haven't reached the detonation point. When detonation point is reached, the solution diverges. Both the point and the time of divergence fits quite nicely the experimental detonation point, so I am quite positive that the divergence happens because of the transition to detonation.

I am running FLUENT's ethylene-air mix as the fluid; standard k-e turbulence model with enhanced wall treatment. The solver is density based. The turbulence-chemistry interaction is Eddy-Dissipation Concept (the only one that works). Inlet boundary conditions are either small velocity inlet (10 m/s, k=0.135, eps=1.718152, T=1500K), or wall with patched hot area. The outlet is either closed or open to atmosphere (pressure outlet).

I tried to change the Flux-Type from Roe-FDS to AUSM, but it doesn't work very well. If I take the data file from few time-steps before divergence, I get divergence right away, and if I start from the beginning, the reaction proceeds really slow.

What would be your advise? What do I do to allow the simulation to run past the detonation point?

Hello Londo!
I am currently trying to simulate DDT in a tube using Methane-Air mixture and was wondering if you ever got to addressing your problem.

No, I didn't. But I stopped working for that company shortly after that post, so it doesn't mean there is no solution.



It is important to mention that to the best of my knowledge methane is very hard to detonate (that's why I used ethylene), so this might be an obstacle as well