[ERPLCFD]

I have been working on a particular problem regarding gaseous film cooling in rocket engine nozzles for a few months now. The majority of my preliminary work is coming to an end (reproducing wind tunnel experiments/ tests conducted at CAL-Tech). However, I have one more preliminary trial to work through before applying my work to a nozzle; I need to test gaseous film injection along a wall.

The current model I have is a planar, 2D duct with a pressure inlet providing air with a total temperature of 646 K and a Mach of 2.44. Further down the length of the duct there is a vertical drop that resembles a step. This "step" is defined with a pressure inlet boundary condition in order to provide flow at a Mach of 1.5 with a total temperature of 312 K. The boundary conditions have been fully defined with the aid of isentropic flow calculations and are not problematic. However, the film I am trying to inject is not the same gas as the "main flow". In short, I would like to vary a number of film characteristics to study the change in film cooling effectiveness. For example, I would like to run the model with a "main gas" of air while the film gas is helium.

The problem I have is frustratingly simple: how do you generate a model with two inlets that inject two different gases into the same domain? I have read many posts on this forum regarding species models vs. multiphase models. I do not believe my situation calls for a multiphase model as both inlets inject gas. On the other hand, species models seem to focus on inlets that inject gases made up of a mixture of other gases. Thoughts?

I have provided a jpeg of the model geometry I am working with.

(Sorry for the long-winded post :-/ )

[ERPLCFD]

I talked to a number of individuals who have experience working with Fluent and managed to figure out the solution to my problem. Selecting the species transport model and inserting various gas species allowed me to inject different gases from two inlets. However, I have run into a new issue: species divergence.

Upon running the model for roughly 1,000 iterations the hydrogen scaled residual abruptly diverges. Luckily the hydrogen residual is the only one that misbehaves, but the reason for the divergence is not known. Thoughts?