Skip to content

tweaks to semiempirical buffet model#18

Open
Asthelen wants to merge 1 commit intoMDOBenchmarks:mainfrom
Asthelen:buffet_cs_fix
Open

tweaks to semiempirical buffet model#18
Asthelen wants to merge 1 commit intoMDOBenchmarks:mainfrom
Asthelen:buffet_cs_fix

Conversation

@Asthelen
Copy link

@Asthelen Asthelen commented Feb 5, 2026

I found that the buffet_openmdao complex step derivatives were being returned as zeros, so in buffet_openmdao I made all np.zero arrays complex which seemed to fix it.

In my case 4 model that integrates this component into the full system, I also had issues with the inputs being arrays vs. scalars, which I tried to resolve by turning inputs into scalars when needed. I also had trouble with the deg2rad function a while ago (don't recall what the error was but I'll try to look into it), so I just manually converted the units using a factor of pi/180.

Also: I noticed that the taper and aspect ratio derivatives are zero for some reason, which perhaps needs to be investigated.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant