Subsonic flow in a supsersonic nozzle
 

Hi there! I am working over a flow problem in a convergent divergent nozzle (convergent section is an x-r profile, minimum throat area is not at a single point but over a section of 20 mm length, divergent is conical) which operates at a chamber pressure of 45 bars and tempreature of approx 3500 K.

I have taken an axisymmetric implicit solver in inviscid case to begin with. Inlet and outlet boundaries are defined as pressure inlet and pressure outlet respectively. Working condition is set to be 0 bar and therefore inlet pressure has been kept at 44 bars instead of 45 (is that right?). Generally, the results have a good match with the solutions of quasi 1D equations, but I have few observations that I can't explain. These include:

1. If I take the air as working fluid, the velocity drops just after the throat and the flow becomes locally subsonic (over a very short length). Velocity however, then suddenly increases and joins the line, generated by the solution of quasi 1D area ratio - mach number relation. Similar kinks are observed for the pressure and temperature cases as well (both increase locally). I know its not an indication of a normal shock wave as in this case, the flow would have remained subsonic over the entire length of divergent section. Does that mean that an oblique shock wave is there? any other idea? Initially I thought that such an occurance is there because minimum area is not at a single point but over a length of 20 mm. So, I changed the profile and observed that no such anomlies are observed, if the minimum area is kept over a single point.

2. If I change the working fluid from air to the gases (gamma = 1.21, molucular weight = 26) which is actually more realistic case than air, I see another velocity drop and temperature and pressure increase just before the divergent sections ends. Such a phenominon was not observed in the case, when air was taken as working fluid. Does that mean that another expansion wave is generatted at the end? any ideas??

I would be grateful for any help, suggestions to improve my results.

Laplacien