fzahle
commented
1 year ago Hi Jelmer, I agree that the current implementation doesn’t really cater to detailed geometry generation, as there is also no inputs for a pre-computed lofted surface which is needed for the computation of structural properties. So we could perhaps consider the definition of blade geometry in a separate section in the yaml, which could allow to directly define the lofted surface as well as the base airfoils needed to generate it.
I can add that I also ran into this issue for the IEA 22MW definition, where I’ve defined a series of airfoils for the the inner part of the blade. So I did as you and interpolated the airfoil data for these sections.
Hi, WindIO currently requires a polar for each airfoil shape you want to provide which means you need many polars when generating an accurate representation of the geometry. At LM we typically define many more cross-sections with airfoil coordinates than we have polars that typically come from wind tunnel testing. We do this for an accurate representation when creating a layup and generation structural properties. To achieve that using windIO format, we would have to interpolate many additional polars, which seem overkill compared to how polars are commonly used in aero-elastic simulation, especially for thinner airfoils.
Could you consider to separate polars from geometry definition? To have the flexibility to define as many cross-sectional shapes (the normalised airfoils) as needed and do with a lot less polars, would be a great benefit.
Happy to discuss!
Best regards, Jelmer