Skin Friction from toal drag testing

[Shayne Young]

I have some highspeed tow data from some testing I have done and want to find out what the skin friction component is from the total drag component. What is the best method for this???

[Ad Hoc]

Depends how much knowledge of naval architecture and hydrodynamics you have and understand. Not to mention the amount of data and its presentation you have too.

[Shayne Young]

The amount of data is small, 2 speeds, 2 surface finishes, 4 resulting loads. This is from modifying some water ski racing ski's, and making sure that I was improving the performance. The result was a resounding yes.
The change that was made, was surface finish.
I'm using the formula (m*g)/(0.5*p*V^2*SA) to get total drag.

Ski Only 60mph 70mph
Drag coeff. Original surface 0.038298171 0.034686226
Drag coeff. painted 0.028686695 0.022593314

Difference in drag of ski only -25% -35%

What I'm finding more difficult is how to extropolate the skin friction component. I have had a quick look at a SNAME paper "skin Friction" and there seems to be a few methods.

Just wandering is there was a commonly accepted method.

Cheers Shayne

[Ad Hoc]

So, you have the wetted surface area of the hull in question and its length? You know its speed?..and you know the water salinity and temperature?

The "total drag" you posted...this is a guesstimate, and not based upon actual testing data...yes? Since you're using a formulae, not actual test tank data.

The skin friction part is easy its is a function of the Reynolds number and the data noted above.

The Cf is then taken from the 1957 ITTC line of: 0.075/(LogRn - 2)^2

Where Rn = Reynolds number.

But this tells you nothing about the finish (and what you expect), it is an assumed value for roughness. You need the total data, ie total drag, but not a calculated an actual measurement. Otherwise all your doing is comparing guesstimates of one against another, of which the ITTC is another too.

You need the total drag and then compare the 2 different cases. But from actual measurements, not formulae. Since if the only difference is the hull finish..ergo..one total drag shall be worse than the other...assuming it makes a whole lot of difference of course. But that is to be determined.

[Shayne Young]

The data I have is actual tow loads, the coefficents are from those loads using the equation stated above.
I use the 1957 ITTC formula to give a good guesstimate on skin friction when wanting to design a new item not built or tested as you mentioned.
I have just never had the finished data and worked the other way.
This is the bit of data I have collected
V - 27.27m/s 31.38m/s
Total drag original surf. - 234.54N 381.38N
Total drag painted surf. - 175.68N 183.28N
Re - 30736222 35375651
Surface Area - 0.15768m2
Length - 1.08m

[tspeer]

Originally Posted by Shayne Young

I have some highspeed tow data from some testing I have done and want to find out what the skin friction component is from the total drag component. What is the best method for this???

I would calculate the drag coefficient or the drag area from the force data. That will eliminate the speed and water density influences. Likewise, calculate the lift coefficient or the lift area.

Then plot the drag coefficient vs lift coefficient squared, and fit a straight line through it. The y-axis intercept is the parasite drag and the slope of the line will give you the lift-induced drag. In reality, there will probably be a lift-dependent component to the profile drag, so not all of the drag due to lift is really induced drag, but the majority of it probably is.

[Ad Hoc]

Originally Posted by Shayne Young

The data I have is actual tow loads, the coefficents are from those loads using the equation stated above.

Sorry for the late reply, been overseas.

You state the WSA as 0.157m^2. This is total WSA. You need to separate out the laminar flow from the turbulent flow, ie need to know where the pins or whatever was used to stimulate turbulent flow is placed on the model and then measure the WSA of each.

You'll also need to know the water temp and viscosity.

All these data should be readily available since the hull was tank tested.

BTW. The velocity seems somewhat high for a model with such a small WSA, as that ship speed or model speed?

[Richard Woods]

As I read it, I think the OP is talking about waterskis.

And that the only changes he has made to the skis in his tests are to the surface coating. He has found a major difference in drag when changing the coating. I think he wants to know why?

Richard Woods of Woods Designs

www.sailingcatamarans.com

[Leo Lazauskas]

For the skin-friction, I found the following coefficients...

U=27.27m/s: ITTC 1957 line = 0.002490, LL08 line = 0.002546
U=31.38m/s: ITTC 1957 line = 0.002436, LL08 line = 0.002499

The LL08 line is my (immodestly-named) variant of Grigson's line.

For an aspect ratio of AR= 0.146/1.08 = 0.135, I found the lift,
induced drag, and Leading-edge suction coefficients:

C_L = 0.2122
C_Di = 0.1061
C_S = 0.1052

I used 40x40 panels in the "Lifting Surface Program" (LSP) which can be found
on boatdesign.net at:
Lifting Surface Program (LSP)

Good luck!

[markdrela]

Originally Posted by Shayne Young

I have some highspeed tow data from some testing I have done and want to find out what the skin friction component is from the total drag component. What is the best method for this???

There is no way to determine the skin friction drag component from the overall drag data.

When you say "friction drag" do you actually mean "profile drag"? They are not the same thing. Profile drag includes friction drag, plus a part of the pressure drag which results from viscous displacement. Both scale closely (but not exactly) with the average skin friction coefficient, so I suppose both could be called friction drag. But in any case, the profile drag can't be separated out from the total drag either.

If you assume that the non-profile drag component (wave+vortex drag) hasn't changed between the two articles, then you can approximately say that
delta(friction_drag) = delta(total_drag)
or better yet,
delta(friction_coefficient) = delta(total_drag_coefficient)
which will remove density and speed variations, as Tom Speer suggested.

[Leo Lazauskas]

Originally Posted by markdrela

There is no way to determine the skin friction drag component from the overall drag data.

When you say "friction drag" do you actually mean "profile drag"? They are not the same thing. Profile drag includes friction drag, plus a part of the pressure drag which results from viscous displacement. Both scale closely (but not exactly) with the average skin friction coefficient, so I suppose both could be called friction drag. But in any case, the profile drag can't be separated out from the total drag either.

If you assume that the non-profile drag component (wave+vortex drag) hasn't changed between the two articles, then you can approximately say that
delta(friction_drag) = delta(total_drag)
or better yet,
delta(friction_coefficient) = delta(total_drag_coefficient)
which will remove density and speed variations, as Tom Speer suggested.

I agree with your (and Tom's) doubts, but it is interesting to try to quantify
some of the components. I think there is some use in deducting the
theoretical values from the experimental drag and seeing what is left over.
Of course, it would be much better if the total drag was available for more speeds.

I can calculate the 2D wave resistance and splash drag with a program I
will soon release here, but I don't know if that is really all that useful for
the low aspect ratio plate under consideration here.

[tspeer]

For a water ski, I would expect the induced drag to be high, due to the small span. At the same time, the speed is high and that will reduce the induced drag.

If we apply some lift-drag bookkeeping, we can divide the drag into parasite drag, induced drag, and wave drag. An empirical approach would be to lump all the drag that is proportional to speed squared into the parasite drag. The induced drag will include all the drag that is inversely proportional to speed squared, but that can be hard to distinguish from the other speed relationships. So an alternative approach would be to call induced drag that component that is proportional to lift squared. Wave drag at high speed can be small, and could be accounted for as either the drag that's left over from the other contributions, or the drag that increases at more than speed squared, or lump it in with the induced drag and neglect it.

The parasite drag will be made up of skin friction and form drag. It's going to be difficult to distinguish between the two. However, given two tests with different surface treatments, the difference can be taken as being all skin friction. The form drag should be similar for the two tests.

All the other drag sources should be similar, too, if the only difference is the surface treatment. So to first order, the difference in total drag is probably a rough estimate of the difference in skin friction. That doesn't tell you the percent change in skin friction, though.

If you really need to know the skin friction as opposed to the parasite drag, then you may have to install some special instrumentation that can measure the shear stress over a patch of the ski's skin. That's obviously a very difficult task for something like a water ski. The only other thing I can think of would be to make a change that would result in a known increase in skin friction. Then you could compare that delta to the overall parasite drag to get the proportion that is due to skin friction.

[Leo Lazauskas]

Originally Posted by tspeer

If we apply some lift-drag bookkeeping, we can divide the drag into parasite drag, induced drag, and wave drag.

Splash drag dominates wave resistance for Fr greater than about 1.1.

Are we sure that these "plates" were towed in the manner of skis. i.e. flat and at AoA?
If so, shouldn't we also try to estimate the wetted length?
Or were they towed edge-on with no AoA?