Convergence problems with non-premixed combustion in a small hydrogen/oxygen rocket engine

jokajoka Member Posts: 1

I'm trying to simulate a non-premixed combustion with gaseous hydrogen and oxygen. I'm only interested in the combustion and mixing of the gases, so my outlet is set to the throat (I thought this would make the case more simple sincethe flow is subsonic only) . As boundary conditions I choosed mass flow inlets (mass flow is fixed) and pressure outlet.

My problem is that I'm not able to get a converged solution, my outlet mass-fraction of h2o is highly oscillating as well as the outlet mass-flow.

For outlet I calculated the pressure with CEA (a program from NASA for rocket engine performance calculations). I'm still confused how to set operating pressure and gauge pressure... I tried to use 0 operating pressure and the calculated value as gauge pressure and also the other way. I also tried to use the average pressure in the flow field as operating pressure.

Maybe you can give me some advice on how to get a reasonable solution.

I use:

Relizeable k-e, energy eqn on, non-premixed, chemical equilibrium, coupled, second order, default and reduced relaxation factors

Thank you very much


  • RKRK Posts: 522Ansys Employee


    Here are a few suggestions that might help :

    1. Pressure inlet boundary condition
    2. K-Omega SST
    3. First turn off the species, let the flow develop in the domain and then turn on the species
    4. If you still see, oscillations, reduce the URF values
    5. Run the simulation with First order until residuals have reduced to a reasonable value and then switch to second order.
    6. Also, try with the pseudo-transient
  • jokajoka Posts: 3Member
    edited June 11

    Hello RK,

    pressure inlet is difficult, because I don´t exactly know the correct inlet pressure conditions to get a proper mass fraction of fuel and oxidizer. What properties should be choosen when turning of species. Some average values of the expected solution, or just air with ideal gas law?

    And is there a basic rule on how to reduce the URF values, for instants wich first, if more than one residual won't converge nicely?

  • jokajoka Posts: 3Member

    Ok, I got a nice converged cold flow, with only species transport turned on. But as soon as I turn on the combustion, my solution starts to diverge... URF values are almost at a minium, the reisudals oscillat around 10^-3 to 10^-4. Do you think it's even possible, to get a better steady state solution, due the highly transient combustion, or do I need to switch to a transient simulation.

  • RKRK Posts: 522Ansys Employee

    When you turn on the species, reduce the URF values for species and see if the oscillations reduce. Also try with pseudo-transient.

Sign In or Register to comment.