Accuracy of adjoint derivatives with actuator disk #524
Unanswered
EdoAlvarezR
asked this question in
Bug
Replies: 1 comment
-
I suggest you check if the totals for targetThrust is accurate. Also, if you want to constrain thrust=drag, you need to create a component like this: dafoam/tests/runTests_MphysAeroPropCoupled.py Line 202 in 234fbd3 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi all,
I have modified the
NACA0012_Airfoil/subsonic
tutorial to include a 2D actuator disk blowing on the airfoil, which works really well as long as the derivative chain doesn't pass through the actuator disk. I have carefully tested my modifications making sure that good convergence is maintained with each modification. The optimizer works nicely with the actuator disk in the simulation, but in the last step of my modifications I coupledtargetThrust
withCD
to match drag and thrust while maximizing CL, which broke the optimizer. The optimality condition is never reduced, which makes me think that the adjoint derivatives get messed up when the derivative chain passes through the actuator disk.What would be the best way of verifying that the partials of the actuator disk are accurate? And, in case they are not, how should I go about fixing them?
Here is the script with all my modifications and I'm attaching a zip file with the full case folder (any feedback on my optimization setup would be appreciated!)
naca0012-subsonic-actuatordisk.zip
Beta Was this translation helpful? Give feedback.
All reactions