Hi, I am running a transient simulation on pulsejet. The outlet is specified as opening, which has only one option---subsonic. However, for pulsejet, velocity can easily go beyond mach 1 at exit(for my case, the max mach number 1.7). We also have compression/expension waves travel back and forth in the tube and combustion chamber. I am just wondering, with supersonic flow at the opening, whether CFX5.7 can model this well, or my results are just wrong.

I specified 'pressure and direction (stable)' at the outlet(opening) with a reletive Pressure=1e5[pa]. My reference pressure is 0[pa]. I use k-omega model.

Currently I am trying to match my simulation results with experimental results. (I monitor pressures at three different points, from which the experimental data are obtained.) The two results looks similar except for the difference at the outlet point. It seems like the reflected expension wave should be stronger or the model has missed a few waves. It it mainly because of the supersonic flow at outlet(opening)?

Other than the outlet boundary conditions, what else should I pay attention to?

Thank you very much!

Hi Jason,

presssure doesn´t travel upstream in supersonic regime so it doesn´t makes sense to set pressure and an supersonic outlet BC would be more realistic. In addition, a finer mesh where you expect waves is desirable (in a steady flow mesh adaption fix the problem) to capture better the waves and even, if the waves are very packed and mesh too coarse, it might be the case, to miss some of them.

Just any ideas, not an expertise oppinion. Hope it helps

pi

Many thanks, pi.

I was also wondering whether it is due to the current coarse mesh. I can import current ccl file to the new mesh, right?

Regarding the outlet, what we want to do is to measure the thrust of the pulsejet, so backflow(if there is) should be considered and measured in the model. That's why I first set it as an opening. However, the simulation results shows we have SOME supersonic flow in each cycle.

Another question is about the thrust calculation. I noticed there is a 'Wall force summary' in CFX-sover. Can I use it to compute the pulsejet thrust and how? Anyone knows?

Regards.