[soleil_du]

Dear all,

I am testing SU2 V3.0 for transonic flows, I found that using Roe scheme gives me a very big drag coefficient while the lift coefficient is ok, and computing with JST (default coefficients) gives a better result but it seems that the numerical dissipation is still so big that the shocks are badly predicted. And it is also the case for a high lift configuration computation.

Does anybody knows why? Which scheme is the best candidate for this kind of application?

Thank you for your help!

[fpalacios]

Originally Posted by soleil_du

Dear all,

I am testing SU2 V3.0 for transonic flows, I found that using Roe scheme gives me a very big drag coefficient while the lift coefficient is ok, and computing with JST (default coefficients) gives a better result but it seems that the numerical dissipation is still so big that the shocks are badly predicted. And it is also the case for a high lift configuration computation.

Does anybody knows why? Which scheme is the best candidate for this kind of application?

Thank you for your help!

Based on our experience. Dealing with inviscid flows... JST and Row perform similar. Dealing with RANS subsonic flows... if you are going to use JST, you should reduce the dissipation AD_COEFF_FLOW= ( 0.15, 0.0, 0.001 ).

Dealing with Roe you typically need slope limiters (2ND_ORDER_LIMITER), Venkat is the best choice, but the selection of the coefficient (LIMITER_COEFF= 0.3) is cases dependent.

Thanks for using SU2,
Cheers,
Francisco