Hi, I am trying to simulate the flow of water out of a nozzle. I created an initial condition with only air in the spray chamber. Despite simplifying the model to the bare minimum I keep on getting the “divergence” message, but it gives me no details as to what the problem may be, only suggestions of things to try.
I ran all kinds of mesh settings, I reduced the relaxation factors, tried different turbulence models, velocity rather than pressure at the inlet. To no avail.
I am sure I must be making a basic mistake somewhere. Can somebody assist please.
Link to simulation
Hi!
Based on the boundary conditions, the velocities within the domain are likely going to be very high. What kind of velocities are you expecting to encounter in this project?
Cheers
I would expect around 13 m/s at the nozzle outlet. I modelled the spray nozzle as a single phase before, and got around 46 cubm/hr flow rate, which has to pass through a 35mm dia nozzle at the narrowest point.
However, I subsequently tried to run at a much lower velocity and had the same problem. I reduced from 235 kPa to 25 Pa - but had the same problem.
Hi!
I have checked the setup and I have some comments:
- Multiphase analysis are transient by nature, and they require additional care during the setup. Since the Courant number is a limitation, we should try to use an optimized mesh, without overly fine cells.
This means a coarse mesh, without layers (or 1 at most). Hex meshes are also preferred, since they tend to capture the interface more accurately.
- When reviewing the initialization of the phase fraction, as well as the inlet boundary condition, we currently only have air (phase 0). Make sure to double check that!
As a side note, if you have a value for the velocity at the inlet, instead of a pressure value, that would be more stable.
- To prevent the simulation from diverging right off the bat, it’s a good idea to set a very small value for the initial timestep size:
For example, 1e-5 or 1e-6 seconds should work better. The timestep will be adjusted accordingly, as the simulation progresses.
Finally, if I may ask, are you looking into a simulation software for your company?
Cheers
Hi, it seems my problem really was the Phase mix-up as you pointed out. The misunderstanding from my side was that I took that value to be an absolute velocity, and did not understand “value” refers to Phase Fraction - it may pay to revise the wording to make that clear.
Thanks for the assistance!
