LES convergence problem:
Hi there:
I am testing fluent LES model using a 3D fully developed pipe flow. The problem is no matter what I did, the mass flow rate at the outlet keeps oscillating, even after thousands of time steps. My testing conditions are summaried as follows: 1. Re = 2500, cylinder length/diameter = 5, periodic conditions are applied. 2. les model: smogrinsky dynamic model 3. a. Numerics: noneiterative timeadvancement + fractionalsstep method or PISO ; Pressure: standard or Presto 3. Initial conditions: I followed the fluent manual. First solved the problem using kepsilon model. Then superimpose the mean velocity of kepsilon as initial conditions for les. 4. grid size: 48*64*128 (r*theta*z) or 64*80*170. All the grids are hex. The equisize skew is less than 0.5. Y plus at the wall is around 0.2. The time step sizes I tried are: 0.002, 0.005, 0.01, 0.05. If you find something wrong with my setup, please let me know. Thanks in advance. Ning 
Re: LES convergence problem:
why periodic conditions are applied?

Re: LES convergence problem:
Because I want to study fully developed pipe flow with small amount of grids.

All times are GMT 4. The time now is 20:10. 