[ex10148]

Hello Everyone!

Case:
Optimisation of front wing of a formula one car using Fluent Adjoint Flow Solver

Issue found:
With the conventional flow solver the residuals tend to converge within 350 iterations using k-epsilon turbulence model with standard wall function. But with the adjoint flow solver it diverges. Issue seems to be in the control settings of the adjoint flow solver. Images are included for the current settings used and the residual history for both the standard flow solver and the fluent adjoint flow solver.

I would really appreciate if any of you could help me where I might be going wrong. Your help will be much appreciated.

Thanks a lot and best regards,
Ahmed.

[alexandrepaes]

Originally Posted by ex10148

Hello Everyone!

Case:
Optimisation of front wing of a formula one car using Fluent Adjoint Flow Solver

Issue found:
With the conventional flow solver the residuals tend to converge within 350 iterations using k-epsilon turbulence model with standard wall function. But with the adjoint flow solver it diverges. Issue seems to be in the control settings of the adjoint flow solver. Images are included for the current settings used and the residual history for both the standard flow solver and the fluent adjoint flow solver.

I would really appreciate if any of you could help me where I might be going wrong. Your help will be much appreciated.

Thanks a lot and best regards,
Ahmed.

Hi, Ahmed

I've been working with fluent's adjoint solver on the last months.

Have you ever solved your problem? I'm dealing with the same issue.

How did you adjust your adjoint control's factors?

[elmo555]

Hey there,

did you ultimately get your calculation to converge?

I'm using the adjoint solver in R17.1, and I fixed my convergence issue by enabling stabilization (Adjoint Solution Controls -> Use Stabilized Scheme) with default settings. Now it converges, but very slowly.

Does anybody have tips on those parameters?