CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   implicit or explicit (http://www.cfd-online.com/Forums/fluent/44894-implicit-explicit.html)

laure June 1, 2007 09:19

implicit or explicit
 
Hello,

I try to model hydrogen-oxygen diffusion in air. I would like to know which solver formulation is better between implicit and explicit.

Thanks

rosol June 1, 2007 09:58

Re: implicit or explicit
 
I think the impicit solver is better and it's more accurate but it take more time.

Razvan June 2, 2007 06:45

Re: implicit or explicit
 
In my opinion, such a problem is evolutive so it should be solved using unsteady formulation. In this case, the theory and the numerical tests say that the explicit unsteady formulation is more accurate then the implicit unsteady.

The implicit formulation is more diffusive ,especially at CFL>1, thus it amplifies the chemical species diffusion. If we add the numerical diffusion induced by the discretisation algorithms and the one generated by the grid, we end up with totally wrong results.

So, I will recommend a carefully constructed hexa-dominant mesh with low skewness and sufficient density in the areas with high species gradients, high order discretisation schemes (at least 2nd order upwind, MUSCL would be better, especially if tri/tets are present), an unsteady explicit solver formulation, with CFL=1 (you can increase it upto 5 ,but some error should be expected). Also, if there are jets, shear layers, or swirling flows involved, you need a very good turbulence model, so RSM is a minimum.

Good luck,

Razvan


All times are GMT -4. The time now is 00:01.