Add center of force calculations #247
Conversation
… chord for the moment computations
…rgence. also fixed the mgpc setup for ank turb ksp
anilyil
requested review from
lamkina
and removed request for
marcomangano and
sseraj
src/solver/surfaceIntegrations.F90
Outdated
| coFz = globalvals(coForceZ:coForceZ+2, :) | ||
|
|
||
| ! before we go any further, compute the actual COP for each time spectral instance. | ||
| do sps=1, nTimeIntervalsSpectral |
There was a problem hiding this comment.
Is there a reason you are doing an extra loop over the time spectral instances to compute the CoF values? It seems like they could be computed in the same loop that starts here.
There was a problem hiding this comment.
The main reason I did this was because similar computations were happening there. To reduce the number of loops, I moved it inside the main loop. Current me also thinks this ended up being better.
|
@lamkina I also added tests for mavgvi. I test function value itself, and the adjoint sensitivity with the database, and the complex step derivative is checked against adjoint derivative in the database. I have noticed we don't really have any adjoint tests for face integrated quantities other than the actuator zone tests. We may want to do a "test_functionals" like test for all face-integrated functionals down the line. This PR is ready for another round. |
|
I think I addressed all of your comments @sseraj. Can you please take another look? |
Purpose
This PR adds the center of force computations as ADflow cost functions. I am calling this center of force and not center of pressure because we look at the sum of all forces, which is what is more relevant for RANS solvers.
The center of force cannot be exactly determined from just the total forces and moments; however, the total moments can be computed from the center of force and total force calculations. However, I kept the moment computations as is and did not modify them. This is because the code already had a very good and granular implementation for the moment terms. It is also a good sanity check to compare the relationships between moment, total force, and center of force.
The center of force is returned in the following functionals:
The first coordinate determines the force direction, and the second coordinate determines the component of the center of force. For example,
cofxyis the y-coordinate of the center of x-force. It is up to the user how these are combined and used.I also added a cost function called
mavgvi, which is just a derived functional based on some compressible relationships. Not useful for most people.Expected time until merged
Similar to #246, after #231 is merged, I need to update this PR. Currently it looks like there are tons of changes but in reality this is a standalone feature. I also need to add tests.
Type of change
Testing
Checklist
flake8andblackto make sure the Python code adheres to PEP-8 and is consistently formattedfprettifyor C/C++ code withclang-formatas applicable