[Foxhunter]

Hello all,

I am simulating combustion in a flare. For this I am using the transient solver, the solution has a hard time converging in the steady state solver. But here is the problem: My energy residual keeps switching between two values. If I change the timestep the values change, but it still switches between two values(so for a timestep of 1e-5s it switches between 1,311e-4 to 3,843e-4 and with a timestep of 1e-6 it switches between 1,423e-5 to 4,562e-5). If I reduce the residual far enough (0,5 or lower) it gets stuck on one value. How do I solve this?
Thanks!

[flotus1]

Are you sure that you are not just seeing the normal convergence behavior of a transient simulation? High residuals at the beginning of a time step that are then reduced to lower values during the inner iterations...

[Foxhunter]

Originally Posted by flotus1

Are you sure that you are not just seeing the normal convergence behavior of a transient simulation? High residuals at the beginning of a time step that are then reduced to lower values during the inner iterations...

Maybe... I'm just wierded out because it starts jumping between two values after some iterations. For each timestep the same values!

[flotus1]

You can upload a screenshot if you want.
Until then I suspect the residuals look like this.

[Foxhunter]

Originally Posted by flotus1

You can upload a screenshot if you want.
Until then I suspect the residuals look like this.

If only...
It converges a little bit, and then it starts changing between the same two values in one timestep. So basically it starts to oscillate between two set values untill the maximum iterations/timestep is reached. And then it does it again for the next timestep.

[flotus1]

A picture is worth a thousand words...

[Foxhunter]

It looks worse than it is. The first few timesteps look like the image you uploaded, it became so irregular after the spark ignition kicked in. I could not replicate the "jumping between two set values", but this is definitly how it looked. What do you think?

[flotus1]

Looks fine to me. The time step size might be too high which causes a high level for the continuity residual and poor convergence in general.

[Foxhunter]

Originally Posted by flotus1

Looks fine to me. The time step size might be too high which causes a high level for the continuity residual and poor convergence in general.

I tried decreasing the time step to 1e-8... Even then it does not converge good. I also changed the relaxation factors, no joy there either. Maybe it's so unstable that normal convergence cannot be achieved?

[Debi]

Hi
I am also facing same type of problem.Did u get the solution!!!
Thanks