Coupled Propeller Movement in Aerostructural Case

[bernardopacini]

Purpose

This PR adds functionality to carry out an aerostructural analysis (and optimization) considering propeller movement. When implemented, the propeller location is tied to the wingbox structure so that it moves with the wing as the wing deflects. This is done by adding structural nodes to the structural mesh, but adding / subtracting fictitious aerodynamic nodes that are used by MELD to transfer loads and displacements between the aerodynamic and structural analyses.

The exact case setup, including mesh generation, is a bit convoluted, so I will add documentation for how to set up this type of problem.

Type of change

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (non-backwards-compatible fix or feature)
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Documentation update
• Other (please describe)

Testing

I have tested that the functionality, and gradients, work as intended, but I have not added automated testing. I can add a test for this before the PR is merged once we finalize the implementation.

Checklist

• I have run unit and regression tests which pass locally with my changes
• I have added new tests that prove my fix is effective or that my feature works
• I have added necessary documentation

[codecov]

Codecov Report

Merging #322 (7560364) into main (3222738) will decrease coverage by 13.43%.
The diff coverage is 100.00%.

❗ Current head 7560364 differs from pull request most recent head d27732e. Consider uploading reports for the commit d27732e to get more accurate results

@@             Coverage Diff             @@
##             main     #322       +/-   ##
===========================================
- Coverage   84.21%   70.78%   -13.44%     
===========================================
  Files         298      250       -48     
  Lines       16330    15100     -1230     
===========================================
- Hits        13752    10688     -3064     
- Misses       2578     4412     +1834     

Impacted Files	Coverage Δ
dafoam/pyDAFoam.py	84.05% <100.00%> (-8.24%)	⬇️
src/include/createAdjointSolid.H	0.00% <0.00%> (-100.00%)	⬇️
src/adjoint/DAObjFunc/DAObjFuncMass.C	0.00% <0.00%> (-100.00%)	⬇️
src/adjoint/DAStateInfo/DAStateInfoLaplacianFoam.C	0.00% <0.00%> (-100.00%)	⬇️
...joint/DAStateInfo/DAStateInfoScalarTransportFoam.C	0.00% <0.00%> (-100.00%)	⬇️
...int/DASolver/DALaplacianFoam/createRefsLaplacian.H	0.00% <0.00%> (-100.00%)	⬇️
...int/DAStateInfo/DAStateInfoSolidDisplacementFoam.C	0.00% <0.00%> (-100.00%)	⬇️
...t/DASolver/DALaplacianFoam/createFieldsLaplacian.H	0.00% <0.00%> (-100.00%)	⬇️
.../DAScalarTransportFoam/createRefsScalarTransport.H	0.00% <0.00%> (-100.00%)	⬇️
...AScalarTransportFoam/createFieldsScalarTransport.H	0.00% <0.00%> (-100.00%)	⬇️
... and 122 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[bernardopacini]

The failed test is due to missing functionality that is needed within MPHYS. A PR is already open to add these updates into MPHYS.

[bernardopacini]

I updated the pre-coupling group to add an error when the warper is requested in addition to masking. This should be ready to merge, though let me know if you have other comments / questions about it. I did not add a test, but I can if you'd like.

[friedenhe]

Great! Can you add some comments at the beginning of the new components and functions? Also, in some components, you have forward mode implementations like: if mode == 'fwd':. I suggest you add a comment there that mphys/dafoam does not support forward mode yet but the code is for future reference.

[friedenhe]

BTW: I think we should a add test for this new feature. Let me know if you need to update the test docker image and what version of dependencies you need.

[bernardopacini]

BTW: I think we should a add test for this new feature. Let me know if you need to update the test docker image and what version of dependencies you need.

Would you prefer a new test, or simply add an actuator disk + movement to the current aerostructural test? In either case we will need an updated wingbox with RBE elements to the actuator disk.

[friedenhe]

BTW: I think we should a add test for this new feature. Let me know if you need to update the test docker image and what version of dependencies you need.

Would you prefer a new test, or simply add an actuator disk + movement to the current aerostructural test? In either case we will need an updated wingbox with RBE elements to the actuator disk.

I suggest we add the act movement to the current aerostruct test.

[bernardopacini]

BTW: I think we should a add test for this new feature. Let me know if you need to update the test docker image and what version of dependencies you need.

Would you prefer a new test, or simply add an actuator disk + movement to the current aerostructural test? In either case we will need an updated wingbox with RBE elements to the actuator disk.

I suggest we add the act movement to the current aerostruct test.

I agree. I will modify the wingbox in the test and add a propeller. Since it is mostly a functionality test I won't refine the aerodynamic mesh for the actuator zone.

[bernardopacini]

I updated the aerostructural case to include the propeller-wing coupling as well as a parent + child FFD setup. I removed the AOA variable to do this as it is more realistic. I also updated the file slightly to match the new MELD and pyGeo versions, so I think we will need to update those in the DAFoam Docker image. I expect the MPHYS tests on the PR to fail now until we update the image as well as the wing BDF file.

[bernardopacini]

This PR is blocked by mdolab/pygeo#144. Once that one is merged we can update the Docker image, verify these tests, and merge in this PR.

[bernardopacini]

The two upstream pyGeo PRs were merged in. I verified the functionality of the propeller motion with 1 and 2 propellers and I updated the MPHYS aerostructural test (sent offline). This PR should be good to go!