Is Matte Faster?

Petros, thank you for your very detailed reply. Much appreciated.
Yes, I could have been more specific about drag, but I also really wanted
to know what accuracy meant for other people and in other contexts.

I still find it amusing that some organisations spent many thousands of
of dollars on some CFD programs that used the ITTC line for the skin friction
(which can be written in one line, or done on a pocket calculator).

I also agree that the predicting the drag of complex shapes made of multiple
components is well nigh on impossible. Good experimentalists will still be
needed for many years. As someone once wrote "A good experiment is
forever".

Addressing an AIAA meeting many years ago, Prof. Peter Bradshaw grimly
predicted that CFD was an accident waiting to happen. I don't know whether
that has come to pass yet, but I can see it happening in a variety of
areas if CFD predictions are not supported by good experiments and
engineers who trust in common sense and experience.

Later this weekend I will post a collection of predictions for the skin-friction
of simple flat plates. I think you will find the scatter between the various
methods quite amusing too. It will also highlight the dilemma facing the
ITTC and other organisations in trying to find a single simple estimating
procedure. Everyone (me included) loves their own method more than those
of their competitors.
And that is for a simple flat plate without all the "real" effects like roughness
rust and rivets, heat or ambient waves, and much more.

All the best,
Leo.

 

Erwan, I have posted some comparisons of various skin-friction lines in
http://www.boatdesign.net/forums/hy...in-friction-lines-some-comparisons-46272.html
I'm not sure if that will help you, or confuse you further.

 

Erwan, what is the purpose of your analysis? Do you need to predict the absolute drag for a particular configuration? Or do you plan to compare drag predictions between configurations to make design decisions? If the latter and the configurations are reasonable similar then the exact choice of skin friction methodology probably won't make a significant difference in the relative results between configurations. What is more important in that case is using appropriate tools for the estimation, being consistent in how each configuration is analyzed, and understanding the probable accuracy of the predictions and the difference between accuracy and precision.

Also, the various skin friction prediction methods are generally straightforward and relatively simple. So you could also estimate the drag using several different candidate methods. Then compare the results, consider your use of the results, and then decide how critical the choice of method is to you for your use.

 

Erwan, David's suggestion is very sensible. Pick a couple of values from
the selection I gave and try them. You will probably find that the final result
will only be a few percent different regardless of which line you choose.

 

I have seen this in other industries as well. It is easy to convince non-technical types that a more complex and costly program will yield better, more accurate results. So these poeple will insist I use this "new" method. When either method does not get you any closer than "five rows of apple trees" (as one of my old supervisors used to say). I remember using an old method to estimate storm water run-off to size a detention tank, it was a simple method that can be done on a slide rule and a few lines of calculations. It was used to size all the storm drains in all the major cities in the USA over the last 100 years, there are some simple "adjustments" that made it reliable and workable. The approving agency made me buy the latest software and redo all the sizing calculations, it integrated an idealized storm run-off calculation over a 24-hour period, because it was more "accurate". I told them it is not more accurate, it only gives you more meaningless numbers to the right of the desimel place. The "new" method had been developed by a Civil engineering collage professor who must have had too much time on his hands. I bought the software and redid the calculations, and it came out within about 2 percent of the old method. And of course they are so confident in the accuracy they will require the tank to be over-sized by double the "estimate". Ignorant people are impressed with complex analysis that they do not understand. One of the reason I am very skeptical of any predictions of the future, in anything from economics to the future climate. the more complex the analysis, the more likely it is not only incorrect, but too easy to hide how it was manipulated to get the desired predictions.

Nice to know things have not changed much in 30 years of very costly research on CFD. One of my engineering professors used to say "one simple experiment is worth a thousand expert opinions". I would also add that it is usually much less costly too. I have seen some very costly "research" where they did sloppy work and did not control for all the variables, so the results are meaningless.

The ignorant are easily swayed by a false and unwarranted faith in our technologists. I think the techno-gurus have become the replacement for soothsayers and crystal ball readers. Ignorant people want to think they are now "sophisticated" by trusting in the pseudo-science of using computers to divine their future doom or fortunes.

This is one of the reasons I have come to call all of my reports full of analysis and complex calculations as "numerology" that divines my client's future. I have become their "guru", and they pay me handsomely for it.

 

The local skin friction, Cfx, is not the same as the total skin friction for the same distance, as the latter includes all the cumulated drag to that point, not just what's going on at one point.

I didn't do a lot of research to pick the best formula, and there is considerable variation in the experimental results. I was simply looking for an approximation that would allow me to pick an appropriate scale for the riblets. If more accuracy was needed, it probably would have been better to use CFD results that accounted for the pressure distribution along the hull and its impact on the boundary layer development.

 

I'm afraid I have really confused the issue in my response to Erwan!
I thought the reference to "Cfx" was to the commercial CFD program of the
same name, and not to the local skin-friction coefficient.

You may well be right about it being more accurate to use a CFD code that
takes into account the pressure effects on BL development, but I'm not yet
convinced. I have read many very recent papers lately and there are many
conflicting results. I'll try to summarise the results and predictions over the
next couple of weeks.

 

Thank you all for providing some "light" on this issue.

Thank you very much Léo for posting all these methods, It's not that important if it confuses me a bit more, it will also prompt me to make some intellectual efforts and it is already very positive. Anyway I will follow your common advice to use different methods. I will take time to digest them and I feel confident, at the end I will have make some progress.

Thanks DCokey, my purpose is to make a brake-down of all drags (aero/hydro & friction/induced) for a typical beach cat.

When it s about "relative value" I do not pay to much attention to the method involved, as it is more a comparison mode, the absolute value is not that important. But to when it comes to establish a hierachy among all drag components, it is more important to have estimation which do not diverge too much from real world.

In fact I try to "reconciliate" figures like an accountant with his asset/liabilities issue.

It starts with real-world datas like righting moment, true-wind speed, boat speed, apparent wind speed and angle.

Then using fluid dynamics formulas, you can compute the driving force.

Under the assumption that at constant speed, driving force = Sum of all drags.
you can then discount step by step all components of drags until you achieve the balance, or I should say the unbalance.

Then you have to cook a little bit the figures within a reasonnable confidence interval to achieve the balance.

But when there is little opportunity for "cooking" induced drag components, the friction coefficient instead seems to be the most fuzzy input.

For the wave drag, I used to interpolate with Serie 64 datas, as the model boat used in the water tank at DTMB fits almost perfectly the size of the beach cat;But of course if hull offsets were available for the beach cat under investigation, I guess I should better use Michlet.

In my former attempt, I had to increase the value of the friction coef by 50%to achieve the balance, do not remenber exactly, do not remenber either which coef I used

This time I wish to use the Excel spreadsheet, to be able to implement new calculations when new datas collected on the water will be available.

Thanks Mr Speer for your remark about Cfx, not sure I would have make the difference

But overall, I love the "spirit" of some comments, which testify of both a large part of experience and wisdom with a little slice of irony. Here I make reference to Petros overview and to Léo's comment regarding million $ CFD program using ITTC.

In conclusion if rocket scientists have such a "distance" with their ultra-sophisticated computer approach it's quite encouraging for a FD benchwarmer like me.

Thanks all for taking time to write these posts

Best regards

EK

 

Using monohull Series 64 data to predict wave drag for a beach cat is likely to result in an error as large or larger than the uncertainty of skin friction drag.

When you say "induced drag" do you mean only the drag due to trailing vorticty from a lifting surface, or do you mean the increase in drag of a lifting surface which is associated with lift, including changes in skin friction?

 

http://www.usna.edu/naoe/people/SCHULTZ PAPERS/Miklosovic, Schultz & Esquivel JoA 2004.pdf would seem to indicate that matte is indeed if not faster, at least more performing at high agles of attack when it comes the the Laser centerbord. The best board was wet sanded at a 60 or 120 grit, which is very rough indeed. A polished board would not perform as well at high lift.

 

From the conclusions in the paper: " The parasite drag coefficient remained constant at 0.007 within the limits of experimental precision. The zero-lift pitching-moment coefficient also remained statistically constant at −0.006. In particular, the reported constant value of CD0 departs from the archive literature, mainly because the scale of roughness in the present work is at least two orders of magnitude smaller at Rq/c ̄≈10−5.

However, roughness effects were present in the parameters af- fected by boundary-layer transition and separation. As a conse- quence, the smoother surfaces performed below the rougher ones. The unsanded stabilizer had a surface finish of 0.46 mils (11.6 μm) and produced the highest observed C L max of 0.969, whereas the polished stabilizer had a surface finish of 0.04 mils (1.0 μm) and produced the lowest observed C L max of 0.808. However, the biggest overall difference in performance occurred between the 60-grit and the 120-grit sanded surface when tested at the higher Reynolds num- ber. This represents a decrease in rms roughness height of 0.03 mils (1.0 μm), yet there was a 13.8% decrease in CL , an 83% increase in CD, a 167% decrease in Cm, and a 53% decrease in L/D for the smoother stabilizer surface finish at a fixed angle of attack of 16 deg."

 

Now let get this straight !!!
what was it sanded with 60Grit WET AND DRY PAPER ??
Or 120 Grit WET AND DRY PAPER ?? THERES A BIG DIFFERANCE IN THE SIZE OF THE SAND MARKS
AND IN WHAT DIRECTION WERE THE SAND MARKS?? ACROSS THE BOARD FROM BACK TO FRONT OR VERTICAL AS IN LONG WAYS TOP TO BOTTOM ??
GOOD INFORMATION JUST NEED TO BE A LITTLE MORE SPACIFIC IN CASE SOME ONE WANTS TO TRY IT !!

 

Read the paper.

 

Here's another one for a flat plate: http://www.usna.edu/naoe/people/SCHULTZ PAPERS/Schultz JFE 2002.pdf

This would be an ideal case for CFD comparison, to study CFDs capacity to predict surface roughness effects.

 

From the data in the paper it appears that at the higher Reynolds Number reported sufficient surface roughness significantly delayed significant separation and the onset of stall. Since the only experimental data reported was the overall CL, CD and CM any discussion of the details of the flow mechanisms involved is ultimately speculative. If similar experiments are undertaken in the future flow visualization, surface pressure measurements, or even field velocity mapping could be very informative.

An interesting comparision would be the same airfoil with a polished surface and boundary layer "trips" of several sizes and in several locations near the leading edge.