problem with v-gutter flowfield
 

hello everyone:

i have been trying to simulate steady flowfield past a v-gutter flameholder for like three months, using both RNG-ke, realizable ke turbulence model. all the spatial discretization scheme are second order. when i use simple scheme for pressure-velocity coupling, the soluting can not converge,the velocity of a monitor point fluctuates. when i use coupled scheme for pressure-velocity coupling, the solution converged, but the length of the recirculation zone is far more larger than what i get from experiments. so, although it can converge, the solution is wrong.

does anybody here has this kind of experiences? could you share your solutiong methods with me?

thand you!

fuxiao