hi,

I am working in an heat sink which is in open atmosphere...... I took the outer domain suffieciently large so that the boundaries should not affect the flow near the heat sink.... The probelm with the simulation are. 1. when coupled solver used, it is not converging....can any one tell me which solver shoud be used for Natural convection problem 2. when segregated solver used ....getting same problem of not converging fully....energy residuals oscilates near 10^-5 and others near 10^-2. 3. I have solved one problem which is fully converged by using coupled solver..... but when i changed my outer domain to large dimension it is not converging i dont know why it is happening.......

Can nyone tell what could be the problem of not convergence

Thanks, Tanveer

Actually Natural convection equations are by nature little difficult to solve. Inertia (convective) terms usually are of lower order than the diffusion terms. And this causes imbalance in the descritized algebraic equation. So any solver for that matter would find it difficult in case of natural convection. To make the things worse, you have the model open to the environment. You might like to mention the Gr / Ra range for which you are trying to solve. For natural convection: convergence of 10^-2/10^-3 is not that poor.

Also try to under-relax continuity and momentum further, may be down to 0.2.

Hope for the best!

thanks rakesh.........

Yes it is very difficult to model open cavity problem in natural convection......now I am doing in 3D and it is taking 4-5 days for convergence.....can you tell how can I confirm the results that it is fully converged........

-tanveer

You can see the Heat Balance report. If it is satisfactory then you may go ahead with post-processing.