convergence problem for super critical fluid accessed from NIST data base

- August 5, 2019 at 6:18 am
  
  ANKITAPUJAHARI
  Subscriber
  
  Hello,
  
  I am using supercritical co2 from NIST database to study natural convection in an enclosure. I have used density-based solver using implicit formulation and the flux type is Roe-FDS. I am starting the simulation from first order. The convergence rate is very slow. What should be done to increase the convergence rate?
  
  Thank You.
- August 5, 2019 at 8:40 am
  
  Amine Ben Hadj Ali
  Ansys Employee
  
  Why density based solver? Any reasons for that?
  
  Smaller time steps could be helpful there. Do not forget to switch on gravity.
- August 5, 2019 at 9:48 am
  
  ANKITAPUJAHARI
  Subscriber
  
  Density-based solver is because of the dependence of density on both temperature and pressure near supercritical temperature and pressure . I have tried using the pressure-based coupled scheme but the probe values are oscillating. One more thing, when I used constant temperature condition on the cold wall convergence resulted. But after using convective boundary condition the convergence rate is extremely slow. I have used steady-state solver