Shape Mismatch for WING_LOCATIONS with NUM_WING_ENGINES = 0

[OscarEspinosaBarcenas]

Hi Aviary Team,

I'm working with Aviary 0.9.8 in WSL and trying to model an aircraft with a single fuselage-mounted engine. I've set the following inputs:

Aircraft.Engine.NUM_ENGINES = 1
Aircraft.Engine.NUM_FUSELAGE_ENGINES = 1
Aircraft.Engine.NUM_WING_ENGINES = 0

I have not explicitly set Aircraft.Engine.WING_LOCATIONS in my inputs, as there are no engines on the wings.

When I run prob.setup(), I encounter this RuntimeError:

RuntimeError: 
Collected errors for problem 'problem3':
  <model> <class AviaryGroup>: The source and target shapes do not match or are ambiguous for the connection '_auto_ivc.v138' to 'pre_mission.core_subsystems.core_mass.fixed_mass.engine.aircraft:engine:wing_locations'. The source shape is (1,) but the target shape is (0,).

It seems that even with NUM_WING_ENGINES = 0 (which should correctly make the target shape for wing_locations be (0,)), an auto-IVC is still attempting to provide a scalar default (source shape (1,)) for wing_locations, leading to the mismatch.

Could you please advise on the recommended way to configure WING_LOCATIONS (or related settings) when NUM_WING_ENGINES = 0 to avoid this shape conflict during prob.setup()?

Thanks,
Oscar

[jkirk5]

Hi Oscar, thanks for reaching out! This error looks familiar, I remember we previously had issues with setting single engines before so this may be related. I will poke around and see if I can figure out where that incorrect shape is getting set by the IVC!

[jkirk5]

@OscarEspinosaBarcenas try updating your Aviary install to at least 0.9.9, that should include the fix that will solve this error. This is the specific fix in question: #709

[OscarEspinosaBarcenas]

Hi Jason,

Thank you for the update. I just installed v0.9.10 and ran the benchmark tests, but several of them failed.

I suspect this might be because I don't have SNOPT installed. Could you confirm if the failing tests are related to that?

For comparison, I ran the same tests on the previous version. You can see the difference in the attached image.

Thanks for your help.

[jkirk5]

I don't think SNOPT is related to any of those tests. At a quick glance those new failed tests all involve reading from data files included with Aviary. Can you provide some printouts/screenshots of the specific errors being raised when you run those tests? You can isolate them easily if you run those files individually as well, outside of testflo