Keel and rudder viscous drag

I have been puzzled for quite a while how to calculate the hydrofoil drag in my VPP. There must be something better than the ITTC 57 correlation line.
Leos Thread http://www.boatdesign.net/forums/hy...in-friction-lines-some-comparisons-46272.html has prompted me to finally write down my thoughts.
Please reed the attached paper, I would appreciate a discussion on this issue.
Uli

 

For full scale prediction I think it would be very important to know the actual transition point and the effect of roughness, waviness and slime. How much does the turbulence due to waves etc affect transition? Are there any measurements about the actual drag and transition points in full scale?

If you look at the data and discussion here, you can see that no actual wing can reach the low values of models: ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19930090976_1993090976.pdf

There have been very different approaches in VPP programs for skin friction for both the hull and the appendages. The effective length used for Cf has had different values, some reduce Cf due to laminar flow for part of the surface and some use roughness allowances.

You mentioned ORC uses high values. Is this from 2012 or from 2013 VPP? They have made a change: "The frictional resistance has also been changed,
using the Hughes frictional line viscous resistance, instead of the classic ITTC as in current VPP." http://www.orc.org/minutes/ITC 2012.pdf

 

Also keep in mind the change in viscous drag of airfoil sections as angle of attack and lift is varied. See any of the drag vs lift plots in Abbot & Van Doenhoff. While this change in drag is sometimes lumped in with induced drag it is a visocus phenomena. The velocity and pressure distribution changes with angle of attack and lift, which affects the boundary layer.

 

Especially interesting are the 60-series foils. Their low drag bucket width changes with Re and roughness, probably also with turbulence level of the wind tunnel, sea or towing tank.

 

Nice paper Uli, thank you. I tend to go a little with David, that for a VPP attempting to predict real boat behaviour, maybe the exact value of Cdo (viscous profile drag at zero lift) is not so important. The lift effects on profile drag, the 3D effects of a relativily low aspect fin, with free surface effects & heel, flow components due to wave orbital motions, sea water contaminations, motions of boat in waves are so much bigger factors. On the other hand, in converting experimental results from model to full scale, a very good understanding of the flow conditions would be essential.

Joakim, do you know how ORCi allows for viscous appendage drag due to lift I don't mean induced drag)? (see also my post http://www.boatdesign.net/forums/hy...ines-some-comparisons-46272-3.html#post617047)

 

The starting point of my thoughts was the fact, that all VPPs that I know use a correlation line for the prediction of keel drag. I wanted to find out, if this is a viable approach. I started with the drag at zero lift. If the correlation line fails already at zero lift, I do not have to investigate the much more complex case with an angle of attack. I think my paper shows that the usage of a correlation line is highly questionable, even at zero lift.

In the first step I limited my analysis to the NACA 0012 wing section. I wanted to avoid the added complexity of the laminar bucket of the NACA-6-series. Again, if the correlation line fails for the NACA 0012, it will for sure also fail for the 6-series.

Concerning roughness and marine growth:
A VPP is only meaningful for the estimation of the best case for a newly polished keel and hull. When I did blue water sailing, I did not spend time to calculate the speed loss caused by marine growth, instead I scrubbed the bottom of my boat.

As for the orbital motion in waves:
Bft. 4 after 8 hours will produce approx. waves with a significant height of 1.16 m and a period of 3.3 sec. The particle speed in the orbital motion will be 2.2 knots and the diameter will be 1.16 m. At the depth of 2 m the diameter will still be 0.55 m. On the other hand, a yacht traveling at 4 knots will need very small vortex generators to trip the boundary layer at the keel. The diameter of such vortices should be approx. 0.3 mm. So the orbital motion is 1000 times too large and will not trip the b. l.

Thanks for the discussion
Uli

 

I agree, that it is most often not sensible to take into account marine growth and excessive roughness in a VPP. However since you are interested in the best available accuracy, it may well be that no actual keel after being in sea for say one day is equal to the accuracy and smoothness of the models used for foil testing. Similar to findings that no actual wing can reach the model performance.

I also agree that the circular motion due to waves will not likely alter the foil behavior, but it will cause velocity gradients in sea and thus also turbulence with many sizes of eddies. It is well know that turbulence tripping depends on the turbulence level of the incoming flow. I don't know the typical turbulence levels in sea. I had a link to turbulence measurements in sea, but it hasn't worked for years.

 

No I don't think it does. It doesn't know the profile so how could it. And at normal leeway angles the profile lift coefficients are small anyway and thus also the change in viscous drag is small.

 

I thought the profile is laser scanned? There's an allowance for thickness (t/c), so why not for lift (CL), it would be proportional to CL^2. I agree, though, that the effect is small for the keel, more meaningful for rudder angles. We are proposing to model rather low aspect, thick wings at an angle of attack with an infinite, zero angle flat plate friction line... I cannot see why not improve the accuracy with known dependencies.

 

Kasper Ljungkvist & co papers on bulb keel wind tunnel tests give a good ground for comparison of various methods http://publications.lib.chalmers.se/publication/148485-shape-optimisation-of-an-integrated-bulb-keel