Precise thickness and volume constraints for non-trapezoidal wings

[sseraj]

Purpose

For non-trapezoidal wings, we use leList and teList with length greater than 2 to define the locations between the root and tip. However, the points currently used in the thickness and volume constraints do not exactly match these intermediate locations. This is documented behavior in DVCon.

As an alternative, I added the option to pass a list for nSpan instead of an integer. Doing so will interpolate within each segment defined by leList and teList and match the intermediate locations. The behavior is unchanged if an integer is passed. I included an example of a cranked arrow wing below.

Thickness constraints with nSpan=10:

Thickness constraints with nSpan=[5, 5]:

The constraints now match the wing break location and follow the chordwise direction.

Expected time until merged

One week

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
• Maintenance update
• Other (please describe)

Testing

I added tests for thickness and volume constraints with a list for nSpan.

Checklist

• I have run flake8 and black to make sure the code adheres to PEP-8 and is consistently formatted
• 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 #127 (2d6d671) into main (2fb4fcf) will increase coverage by 0.04%.
The diff coverage is 90.32%.

@@            Coverage Diff             @@
##             main     #127      +/-   ##
==========================================
+ Coverage   62.56%   62.60%   +0.04%     
==========================================
  Files          43       43              
  Lines       11186    11207      +21     
==========================================
+ Hits         6998     7016      +18     
- Misses       4188     4191       +3     

Impacted Files	Coverage Δ
pygeo/constraints/DVCon.py	70.83% <90.32%> (+0.47%)	⬆️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[marcomangano]

Looks good! I just have one comment on the inputs, but it's not related to the specific additions of this PR

[anilyil]

The changes look good. I only have one minor comment on the coding style regarding the different uses of the nSpan variable. I left it to one of the routines you added but obviously applies to both. Let me know what you think and we can either slightly modify it or I can just approve it if you don't want to make that change.

[anilyil]

I am not sure if I like this change to what nSpan is mid-way through the routine. Initially its the list of spanwise section counts, but then it is switched to the total count. I guess it would be a bit better if we used a different variable for the total (or the span list)? I am not sure how useful my comment is but thought this type of using the same variable name for two things might be confusing?

[sseraj]

I agree that the current approach is not very clean. I can propose two alternatives that would keep nSpan as an integer:

1. Add a separate kwarg that is a list defining the fraction of nSpan that gets allocated to each segment, e.g. [0.5, 0.5].
2. Don't add any additional arguments. For a given uniform distribution in parametric space, 'snap' the closest point to the intermediate location and resample the curves. This would change the default behavior.

[sseraj]

Turns out Anil was referring to the variable names in the function, not the API. I will update the variable names to be clearer.