[Simon Smokes]

Dear All!

My Name is Simon and I'm currently working on my mean flow solver for a SST-turbulence model. In order to ensure numerical stability I used the SIMPLER algorithm (pressure based) and in my momentum equation TVD by applying deferred correction (DC). My problem is related to the boundary condition of the first node referring to the 'inlet flux'.


#node1: u_mom = source_term/a_P

=> source_term = friction_losses + DC + inlet_flux * inlet_velocity
=> inlet_flux = rho * A_inlet * u_inlet
=> inlet_velocity = u_inlet


With the first initial guess of the velocity field I obtain the 'momentum velocity field u_mom' which is reduced by the friction losses represented by the according source term. Due to the momentum residual of 0.0013 the loop is run a second time to further correct my velocity field.


When I go now in the second run of the DC loop the first node of my 'limited velocity field' and obtain the new momentum velocity, my first node differs from the others and oscillation and divergence starts.

Numerical valuesn (example):

u* = [1 1 1 1 1];
u_mom(1) = [0.8 0.8 0.8 0.8 0.8]
u_mom(2) = [0.8 0.6 0.6 0.6 0.6]

The problems araise from the inlet condition: a fixed value for inlet flux and velocity cause problems as well as setting the inlet velocity equal to u_mom(1).

Since years I work on CFD by applying various techniques and I would really appreciate an answer to my problem and support you as well in any case!