Subject: Re: [harryproa] Asymmetric Bi-directional Rudders
From: Rick Willoughby
Date: 8/4/2010, 10:22 AM
To: harryproa@yahoogroups.com.au
Reply-to:
harryproa@yahoogroups.com.au

 

Dennis

The value of analytical models is that they are based on the fundamental physics.  In this regard Leo Lazauskas and I have something in common.  As long as the assumptions made in the model are applicable to case being analysed then you can be led in interesting directions.  Leo can entertain me for hours with his mathematics.

The human brain has the ability to collate the interaction of no more than about three  independent variables.  The power of good analytical models is that they can bring all of the variables together and you can test way outside current experience with some confidence.  Many of the so-called experts on the BD forum simply do not appreciate this and are beyond learning.  I have even seen one fruit-cake argue with Leo that he was not wrong because he was taking it straight from his 20 year old lecture notes.  Have a look at post #459 here:
http://www.boatdesign.net/forums/projects-proposals/windmill-wind-turbine-powered-boats-how-many-out-there-they-viable-14182-31.html#post302445
This guy never makes a mistake - his lecturer was at fault.  He has done no real world testing and learning to verify the stuff he has been spoon fed and is happy to repeat as gospel.
 
The modelling I do applies reasonably well to any size boat and speed.  The Re# is a key consideration.  At low Re# the viscous drag determination gets less accurate.  At high speed you need to be mindful of cavitation.  Javafoil is fine for water foils but it is primarily set up for air.  It has corrections for Mach number but nothing for cavitation so you have to be aware of when cavitation is likely and check separately for it.

If you have a good analytical model you do not need a lot of data.  You only need a couple of points to verify that you can predict with confidence.  You only need to look at gross outcomes to check that you are close.  Believe it or not I have a VPP for all of my pedal boats.  I can determine the required power level to 1 or 2% over a wide range of speed for any particular displacement.  They combine Javafoil data for the prop and Flotilla data for the hull to closely model the boat and its propulsion.  You may think this is trivial compared with a sailing boat but a propeller is far more complex than a sail and I have an analytical model for that.

You need to work through one iteration of the forces acting on your boat to get an appreciation of the scope.  The little model you did on the heeling is maybe 10% of what the overall VPP would look like.  You are then in a good position to start building an analytical model of your boat.   A good first case is close hauled at the speed you hope to achieve with the safety margin you aim to work at regarding distribution of weight. Michlet is good enough for the hulls and Javafoil can do the sails and rudders.  I expect I will be able to show you simple ways to combine them into a complete boat model so you can go sailing at your desk.  It saves getting wet while experimenting.

Rick
On 04/08/2010, at 11:29 PM, Dennis Cox wrote:

 

Rick,
 
I want to say... I really enjoy the way you let us "see" you think.  I have seen you come up with a speculation (not quite a theory yet) on several occasions and talk about it.  Then you go back and do some analysis to bring it up to a theory with some projections.  You have, on occasion, even changed your mind from when it was at the speculation level.  It says a lot about you.  Many try to sweep under the carpet that they had a turn around.  You're a true scientist.
 
Context - With your extensive background with the human powered speed record machines... 
Some theories you had may pay off with big gains in performance.  But to reach the success levels you have, many things must have come with very small gains.  For the example at hand... given a test Proa in the size you usually have (I know yours are not Proas) and changing from NACA0012 that has to rotate around on every shunt to a symmetric NACA0012-DE and on to one of these NACA07-DE designs. 
 
Question - How would you go about quantifying the results and hopeful gains.  In my old career, I'd specify about a million dollars worth of transducers, strain gauges and computers and the technicians would knock themselves out complying.  I'm at a little loss imaging how I could say, detect and document that one ran at 1 degree less AOA and had 5% less drag.  Especially when wind and water conditions will certainly change more than the test results spread from one test to the next. 
 
Thanks,
Dennis

From: willoughby_rick <rickwill@bigpond.net.au>
To: harryproa@yahoogroups.com.au
Sent: Wed, August 4, 2010 5:26:27 AM
Subject: [harryproa] Asymmetric Bi-directional Rudders

 

I have taken a look at what should be the best section for an asymmetric bidirectional rudder.

It threw up a few surprises. The best I could come up is a double ended version of a NACA 07 series:
http://www.rickwill.bigpondhosting.com/Double_End_NACA07.png

At an Re# of 10E6 and aspect ratio of 4 it has an L/D of 27. It has flow separation on the trailing edge at all angle of operation that reduces at higher Re#. But it is no more than 3% of the chord in the normal operating range. The lowest Cd is 0.004.

By comparison a NACA0012 section at same Re# and AR has a best L/D of 17.7. The in-line Cd is actually higher than the asymmetric section at 0.009. One advantage is that it does not have flow separation until about 7 degrees AoA. It would be designed to work at about 4 degrees for nest L/D.

What surprised me is that the large diameter nose actually has better lift, less flow separation and lower minimum Cd than the pointy nose/tail version. The reason becomes apparent when you look at the pressure profile for the respective sections. Notably the 07 series was designed for a flat pressure profile. The NACA 16 series was a development that factored in air compression for higher speed subsonic flight. However not a factor for water although cavitation could need consideration.

This idea has a lot of upside with the one downside of flow separation and possible resulting vibration. There could be some benefit roughing up the nose/tail to help the flow stay attached - could be tested with a strip of sandpaper glued in place down the edges.

Rick



Rick Willoughby
03 9796 2415
0419 104 821

__._,_.___
.

__,_._,___