Burner simulation
 

Hi,

I'm trying to simulate a non-premixed burner. I have run the simulation cold until a convergence of 10e-05. I then run the reacting flow. The simulation runs well for a while, approximately 100 iterations and then starts to oscilate badly. I have set a small under-relaxation factors for the reactives and the density, 0.9. Is this too large?

Any ideas on how to fix this?

Cheers,

Mark.

Re: Burner simulation
 

Not knowing anything about your reaction model etc, I would suggest to use values of 0.5 initially for the reacting components and 0.8 for energy. Later you can slowly increase the values to 0.8 or 0.9 for reaction components and 0.95 or so for energy. I have not noticed a strong influence of the density URF. Make sure your temperature field is initialized at 1000 K or so to get it all going.

Re: Burner simulation
 

Thanks Erwin,

I will do that. I initialised my simulation with temp=1300k and ch4=0.2 to get ignition.

Does this sound ok to you?

Mark

Re: Burner simulation
 

Don't use URF values that are too low, the solution will take 100 years to converge. Use at least 0.5 but don't go too quickly to 0.9.

Re: Burner simulation
 

Thanks Erwin,

Just to confirm, the URF's just ease the change from the old value, say Vn, and the new value Vn+1. Is this correct?

If so, changing the URF's in mid-calculation makes the calculation faster but could also cause the solution to diverge if the change is too sevre?

It's basically just a tool to get convergence with the side effect of longer computation time - am I right?

Mark.

Re: Burner simulation
 

Yes, yup, and sure!

Re: Burner simulation
 

Thanks!!!