I am working in a compressor stage, I am trying to validate the model with results of other software CFX, so replicated the whole model in terms of the geometry and boundary conditions as well initial conditions, though after a few iterations the simulations stops with the following error message "Error: divergence detected in AMG solver: temperature" I check everything over and over but seems that the problem is not there, though I noticed temperature and pressure limited in the volumes of my model. So I wonder if anyone could help me out on this in order to figure it out.

Thanks in advance.

German

You cannot setup CFX and Fluent absolutely identically. There are under-the-hood differences.

So there are differences beyond my control, because of the code of Fluent?

Yup.

Hey

Try lowering the under-relaxation Factor for energy to 0.5 as well as density and see if it works.

Regards

AAA

Never lower energy to anything less than 0.98. Try disabling secondary temperature gradients, fmg initialisation, pressure based coupled solver with low CFL.

Hi Again

Red lemon is right. The energy under-relaxation factor must be between 0.95 to 1.0. What I meant was to lower it for a while to see if an initial solution can be reached, then bump it back to 1.0 for the final solution.

Sorry

AAA

How I disable the secondary temperature gradients?

Type this into the window : (rpsetvar 'temperature/secondary-gradient? #f)

Under which menu?, because I typed into the window, but does not recognize the command.

Type the entire line including brackets into the Fluent windows and hit return. This will disable them and make the solution more stable.