Szymon Szyszko asked 4 years ago
Hi Andre,
I’m currently trying to build my confidence around the method that flow5 uses for aero analysis. For the time being I keep comparing the results with Jibe2 wind tunnel test. I have run many calculations with different analysis setup/meshes etc. Things I have noticed and cannot really fully understand why it happens are:
- As you have covered it in one of your videos, the pitching moment for T6/1 TruUniform-thickSurf analysis behaves oddly in the proximity to alfa=1.5deg due to wake intersection. I have noticed that the problem disappears when thin panels are selected, viscous drag from fuselage is deselected and the analysis is run as T1 analysis. In theory (at least the way I understand it) this should not have noticeable effect on the behaviour of Cm at the difficult AoA. Is that an expected behaviour? If so, why does it happen this way and is it “universally” more reliable?
- In many runs, I have noticed thick surfaces are less reliable than thin panels. Where thin surface behaves quite consistent for different meshes, thick surfaces can give slightly better results or significantly worse. I know that fuselage-wing interaction is more significant with thick surfaces, but I could not find a reliable way of working with that. Do you have any recommendations in the setup of mesh or analysis to benefit from thick surfaces?
Thanks in advance for your help! If any of the topics or its explanation would be covered in detail by some academic papers, I would appreciate all reading recommendations! Also, if it would be easier to cover it on the screen share session, I would be up for that!
Cheers,
Szymon
Szymon Szyszko answered 4 years ago
Thanks Andre! This is very insightful! I send you my project file later over the weekend. Thank you for all your swift responses and support!
Hi Szymon,
1. The problem disappears in the case of T1 or T2 analyses because the wake panels are aligned with the x-axis whatever the aoa. This is the usual simplifying assumptions for VLM and panel methods, e.g. xflr5 or AVL. This enables to solve the linear system only once, since the influence matrix is the same for all aoa, and makes for a fast analysis. The possible intersection of wake panels with the plane needs to be checked only once when the plane is constructed.
The T6 polars are intended to be all-purpose non-linear, with the matrix being built for each aoa. This allows the wake panels to be aligned with the free stream flow for each aoa. The drawback is that the wake intersection needs to be checked for each aoa.
However my recommendation is to use the VPW method which solves elegantly these issues at the cost of longer analysis times.
2. I haven’t noticed this, but then I tend to work only with thick surfaces. Could you please send me a project file which shows these problems?
Academic papers: I’ve listed all the specific references in the Documentation and resources section: https://flow5.tech/index.php/flow5/sample-page/documentation-and-resources/
But my favourite general reference guide is Mark Drela’s “Flight Vehicle Aerodynamics” which you can purchase online.
Regards,
André