Try to change the underrelaxation factors.

Franco

Maybe lowering the CFL number at the beginning of the simulation will helps.

Dear Zun Cai ..

Can you guide me how to change CFL number ? because there is no option appear this kind of simulation. I mean, when we adopted PHASE COUPLED SIMPLE scheme.

Thank you..

Respected Professor Franco Concli.

1) Can you suggest should we decrease the (underrelaxation factors) URF or increase ?

2) For divergence in Epsilon which URF is more important among ( pressure, density, body force, momentum, volume fraction, turbulent kinetic energy, turbulent dissipation rate, turbulent viscosity, air bin ) ?

Please guide..

Thank you for your answers.

The model is axisymmetric and I´m using default under-relaxation factors, but will try to change parameters.

Hi Daniel Carlos Andrade,

How good is the mesh ? If the mesh is good, try running the simulation with first order scheme for some iterations. Once the residuals are flat switch it to second order. Also reduce the URF/CFL number.

Usually, convergence issue is due to bad quality mesh, unrealistic boundary conditions, complex physics etc.

There are many ways to deal with this issue, like:

-reducing URF/CFL number

-using first order schemes initially and then moving to secondary order.

-using pseudo-transient schemes etc.

Hope this help you.

Thanks & Regards,

Rajesh.

Hi Rajesh Alayil,

The mesh is 100% orthogonality, aspect ratio is almost 15, and I´ve refined in relevants zones. I´m using 2D geometry. However, epsilon still diverges.