[Kepler12]

I am trying to simulate the flow through a simple converging-diverging nozzle with known inlet and outlet pressure and temperature.

Inlet
Pressure: 700000 psi
Temperature: 3400K

Outlet
Pressure: 101325 psi
Temperature: 300K

It is 2-D density based and axisymmetric using air as the fluid. Energy is on and I am using the k-omega SST viscous model.

I am not sure what I am doing wrong but the simulation consistently diverges. Below I have attached 1) Scaled residuals chart, 2) Best Mach Contour number produced, and 3) Mesh Quality Stats, all for ~225 iterations.

I am fairly new to CFD, so feedback would be greatly appreciated.

[Indra]

Originally Posted by Kepler12

I am trying to simulate the flow through a simple converging-diverging nozzle with known inlet and outlet pressure and temperature.

Inlet
Pressure: 700000 psi
Temperature: 3400K

Outlet
Pressure: 101325 psi
Temperature: 300K

It is 2-D density based and axisymmetric using air as the fluid. Energy is on and I am using the k-omega SST viscous model.

I am not sure what I am doing wrong but the simulation consistently diverges. Below I have attached 1) Scaled residuals chart, 2) Best Mach Contour number produced, and 3) Mesh Quality Stats, all for ~225 iterations.

I am fairly new to CFD, so feedback would be greatly appreciated.

Can you send a snapshot of solution methods. There's a lot of things that can be don eher to imporve convergence. the first thing i would do is change to first order schemes. Also, are you getting any messages in console durng the run? That could tell us what relaxation factirs to reduce.

[Kepler12]

Thanks for the help Indra.

I attached a screenshot of my solution methods. For the next iteration I will change them all to "first order upwind". Just for future understanding, how might changing the solution method to first order allow for convergence?

In the console the message "time step reduced in X cells due to excessive temperature change" after every few iterations. X being some number usually between 2-15. I did not want to attach a screenshot of my console because it has some sensitive information. I've also got a message saying "temperature limited in X cells.

Edit: still divergence while using first order equations.

[Indra]

Originally Posted by Kepler12

Thanks for the help Indra.

I attached a screenshot of my solution methods. For the next iteration I will change them all to "first order upwind". Just for future understanding, how might changing the solution method to first order allow for convergence?

In the console the message "time step reduced in X cells due to excessive temperature change" after every few iterations. X being some number usually between 2-15. I did not want to attach a screenshot of my console because it has some sensitive information. I've also got a message saying "temperature limited in X cells.

Edit: still divergence while using first order equations.

Right, there we go, so you have to go to solutions>controls>limits and reduce the positivity rate. This will make sure that the rapid temperature change is controlled and you shouldn't get any report sayig temp. limited.

Also, I have seen that AUSM scheme in under flux type is more stable than ROE-Fds for mach number>1.5.

As for your questions regarding schemes, first order increases the convergence but it is not as accuracte as second order. So we usually run enough iterations in first order to make sure our residuals drop and then swith to second order. There's also a method to use blended first and second order approach which you can access through the TUI but again, don't worry about it for now. Always change oe thing at a time. For now try changing positivity rate and switch to first order for residuals to drop.

[MahbubTariq]

Hi Kepler 12,
Have you been able to sovle the convergence issue? I am facing similar issue while modeling a convergent nozzle. I am having stuck residuals in the order of 10e-1.

Thanks.