Simulations of wing-like structures always have detached flow

[tdunning]

In trying to simulate a crude wing in 2-D, I always seem to have flow that detaches from the top of the wing. This happens with low and high reynolds numbers and with low and high viscosities and very low angle of attack.

Could you help me build a simple example that does this correctly?

Here is an example:

[weymouth]

This depends sensitively on the foil shape and on the resolution of the flow. I suggest implementing a real foil shape and then reproducing some test cases from previous research papers. Here's one example, but there are lots and lots.

[weymouth]

Here's the preprint of the paper. https://eprints.soton.ac.uk/369635/1/Maertens2014%2520CMAME.pdf

[weymouth]

I added some code a few days ago to https://github.com/weymouth/WaterLily.jl/issues/53#issuecomment-1554818303 which can be immediately used to embed a parametric foil geometry, such as a NACA.

curve(t::T,thk) where T = SA[t^2,T(5thk*(0.2969t-0.126t^2-0.3516t^4+0.2843t^6-0.1036t^8))]
NACA_sdf(thk=0.12) = parametric_sdf(t->curve(t,thk))

function test()
    sdf(x,y) = NACA_sdf()(SA[x,y])
    x = range(-0.5, 2.5, length=100)
    y = range(0, 0.5, length=50)
    z = sdf.(x',y)
    contour(x,y,z,aspect_ratio=:equal, legend=false, border=:none)
    contour!(x,y,z,aspect_ratio=:equal, legend=false, border=:none,levels=[0],color=:green)
    savefig(".\\figures\\distance function.png")
end

Gives this figure:

[tdunning]

This looks very promising

On Mon, May 22, 2023, 03:39 Gabriel Weymouth @.***> wrote:

I added some code a few days ago to #53 (comment) https://github.com/weymouth/WaterLily.jl/issues/53#issuecomment-1554818303 which can be immediately used to embed a parametric foil geometry, such as a NACA.

curve(t::T,thk) where T = SA[t^2,T(5thk*(0.2969t-0.126t^2-0.3516t^4+0.2843t^6-0.1036t^8))]NACA_sdf(thk=0.12) = parametric_sdf(t->curve(t,thk)) function test() sdf(x,y) = NACA_sdf()(SA[x,y]) x = range(-0.5, 2.5, length=100) y = range(0, 0.5, length=50) z = sdf.(x',y) contour(x,y,z,aspect_ratio=:equal, legend=false, border=:none) contour!(x,y,z,aspect_ratio=:equal, legend=false, border=:none,levels=[0],color=:green) savefig(".\figures\distance function.png")end

Gives this figure: [image: distance function] https://user-images.githubusercontent.com/468263/239883535-af035c41-e03e-4c75-9935-e05595b56c3d.png

— Reply to this email directly, view it on GitHub https://github.com/weymouth/WaterLily.jl/issues/56#issuecomment-1556983042, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAB5E6WJ5PKU5NBFLKLJX5DXHM67FANCNFSM6AAAAAAX75UVH4 . You are receiving this because you authored the thread.Message ID: @.***>

[weymouth]

Note that 2D simulations at low Re should give separated flow.

High Re flow is always 3D, if you don't run a 3D simulation at high Re, you will get something like this

Attached, but weird.

Gabriel D Weymouth