CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   SU2 Shape Design (https://www.cfd-online.com/Forums/su2-shape-design/)
-   -   Continuous adjoint converge issue in transonic airfoil shape design (https://www.cfd-online.com/Forums/su2-shape-design/222944-continuous-adjoint-converge-issue-transonic-airfoil-shape-design.html)

zloty001 December 15, 2019 10:13

Continuous adjoint converge issue in transonic airfoil shape design
 
Hello everyone !

I am dealing with shape design problem and I have got some questions about adjoint parameters, thus I have troubles with converge adjoint calculations.
My direct case was converged well.
  1. What would make adjoint soultion converge and do it quickly ? I have been changing frozen turbulent viscosity, removing sharp edges and changing CFL_REDUCTION_ADJFLOW, LIMIT_ADJFLOW and CFL_REDUCTION_ADJTURB and it didn't bring me good results.
  2. How should I understand CFL_REDUCTION_ADJ parameters? The value I give is reducing CFL like: CFL_NUMBER * CFL_REDUCTION_ADJ = Adjoint CFL number ? Could someone provide me a good guess what factors should I use ? I've tried few combinations in range 0.001 to 1 and even over 1 but none was proper one.
  3. Should I use FROZEN_LIMITER_DISC, FROZEN_VISC_DISC, ADJ_SHARP_LIMITER_COEFF or any of those ADJ parameters? Or should I use only ADJ parameters for flow, turbulence and linear solver ?

To the flow solution I use RANS solver with SA turbulence model, Mach number is 0.85 and Re= 4.6e6.
Direct calculation converged well with CLF number = 50, FGMRES linear solver with LU_SGS preconditioner, entropy fix coefficient = 0.5, second order flow equation and first order turbulence equation.

I attach config file, mesh, screenshots, solutions and restart solution files in link below. I got stuck on this problem for several weeks and would appreciate if someone could help me.

https://github.com/zloty001/CRM_ShapeDesign

Regards,
zloty


All times are GMT -4. The time now is 05:59.