CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Multiphase flow and convergence issues (

Erik October 13, 2007 22:37

Multiphase flow and convergence issues

I am trying to analyze a liquid rocket injector that injects liquid oxygen radially outward from the center of a combustion chamber into a cylindrical sheet of liquid kerosene that flow axially down the outside of the injector (so the lox impinges the fuel at a 90 degree angle).

My first question has to do with reversed flow. I'm rather new to using Fluent so I've started by setting up a "simple" water flow model. I've taken a 30 degree pie slice of the injector (the injector elements are symmetric every 30 degrees) and I'm running with water. I'm not modeling the chamber; my geometry captures the injector elements and a short distance out past the location where the fuel and lox impinge each other (so my inflows are the injector elements, I have symmetry planes on each side (which bound the symmetric pie slice) and then I have outflows on the top, back, and bottom. I'm using a basic k-epsilon viscosity model and I've activated the energy equation so I can model the water at the expected lox and fuel temperatures (90 and 300K respectively). When I try to run this model I don't any iteration data in the command window and it doesn't appear to converge (even after 15,000 iterations). All I see in the command window is "reversed flow in xxxxx faces on outflow 9", where "xxxxxx" changes each iteration. Any suggestions?

Secondly, how does one model a cryogenic fluid in Fluent? Will this require the use of a UDF? Would it be possible for someone to outline the steps I need to take to set up a multiphase VOF model assuming I will have liquid oxygen, gaseous oxygen, and liquid kerosene? It's not obvious to me how I specify that oxygen is in liquid form and how I incorporate that into my inflow boundary conditions.

What I suspect happens is that the liquid oxygen is injected as a liquid but almost immediately upon leaving the injector begins to vaporize. So we have oxygen (gas and liquid) impinging a kerosene fuel sheet which then causes the kerosene to separate into droplets, further lox to vaporize, and mixing of the kerosene and oxygen.

I know this is not a simple problem so please feel free to ask any additional questions if you require clarification on what I'm trying to do.

Thank you!


eseller July 13, 2009 17:08

same problem
Hi Erik,

I was curious on the progress you made with your rocket injector model? I am currently taking on a very similar task to your past one. I am modeling LOX/methane combustion in a liquid rocket motor and am having difficulty obtaining convergence.

Some differences from your model are that I am dealing wiht a 1/16 slice of the injector/chamber/nozzle and I am trying to model the propellants as high-density gases as liquid droplets/vaporization require too much memory. With your experience, what is your opinion of this strategy (density-trick)?

I am being very general, but would like to know first if you or anyone else is able to offer up some help and then I can go ahead an layout my setup in Fluent and ask for specific advice concerning inputs and specific parameters.

Thank you!

--EriK (it's always nice to see the name spelled correctly :cool:)

All times are GMT -4. The time now is 18:26.