[wruudf]

When I'm trying to use HLLC as the convective numerical method to run the NACA0012 test case in quick start, the residual keep a very high level and never converged.

Like this:

Iter Time(s) Res[Rho] Res[RhoE] CLift(Total) CDrag(Total)
3280 0.197821 -3.372983 2.177480 0.335392 0.022200
3281 0.197818 -3.441973 2.102265 0.335399 0.022206
3282 0.197817 -3.541882 1.945486 0.335422 0.022213
3283 0.197816 -3.534004 2.001859 0.335395 0.022203
3284 0.197815 -3.524070 2.163939 0.335392 0.022202
3285 0.197813 -3.515924 2.023638 0.335435 0.022217
3286 0.197811 -3.554574 1.965987 0.335365 0.022194
3287 0.197809 -3.339415 2.204189 0.335412 0.022212

Roe seems to have the same result.
1. Are these two method still available in SU2? and how can I make them work?

Also, I've checked the numerics_direct_mean.cpp, and it does have some code about HLLC and Roe method.
2. Is this file the right place if I want to make a change on the solvers?

[hlk]

Originally Posted by wruudf

When I'm trying to use HLLC as the convective numerical method to run the NACA0012 test case in quick start, the residual keep a very high level and never converged.

Like this:

Iter Time(s) Res[Rho] Res[RhoE] CLift(Total) CDrag(Total)
3280 0.197821 -3.372983 2.177480 0.335392 0.022200
3281 0.197818 -3.441973 2.102265 0.335399 0.022206
3282 0.197817 -3.541882 1.945486 0.335422 0.022213
3283 0.197816 -3.534004 2.001859 0.335395 0.022203
3284 0.197815 -3.524070 2.163939 0.335392 0.022202
3285 0.197813 -3.515924 2.023638 0.335435 0.022217
3286 0.197811 -3.554574 1.965987 0.335365 0.022194
3287 0.197809 -3.339415 2.204189 0.335412 0.022212

Roe seems to have the same result.
1. Are these two method still available in SU2? and how can I make them work?

Also, I've checked the numerics_direct_mean.cpp, and it does have some code about HLLC and Roe method.
2. Is this file the right place if I want to make a change on the solvers?

Thank you for your question
1. These methods are still available in SU2. For all solvers, whether and how quickly it converges is sensitive to many factors - the solver settings, the mesh, and the flow conditions all effect this. The settings in the QuickStart should be tuned to work well with the solver specified in that file - changing the solver may require things like lowering the CFL number or using the limiters. You will find more examples in the TestCases folder available on github. Make sure that you check out the version that corresponds to the version of the code you are using.

2. yes.

[wruudf]

Thank you very much. I just find there is a NACA0012 in incomp_euler folder and it works with Roe. But is there any way that I can get the description of this test case instead of reading all the information from the cfg file?

Originally Posted by hlk

Thank you for your question
1. These methods are still available in SU2. For all solvers, whether and how quickly it converges is sensitive to many factors - the solver settings, the mesh, and the flow conditions all effect this. The settings in the QuickStart should be tuned to work well with the solver specified in that file - changing the solver may require things like lowering the CFL number or using the limiters. You will find more examples in the TestCases folder available on github. Make sure that you check out the version that corresponds to the version of the code you are using.

2. yes.