Appropriate Scale for Modeling

[fcfc]

See Design project - model tow test

Seems you need at least a 2m length model (7 ft).

[daiquiri]

Originally Posted by MastMonkey

I accept that I know nothing about fluid dynamics, but I have seen numerous example of models being tested for aerodynamic and hydrodynamic behaviour and some of them have been quite small in comparison to the full sized object.
I assumed that there was some scaling factor that was applicable to these models that was determined by a some principle of FD. I guess there isn't.

Let me be frank - that's a VERY bad news, particularly the bolded part.

You absolutely need to have a good grasp in fluid dynamics and in physics if you want to get any meaningful result from your testings, or even just to know in what direction your research should go.

There has been a lots of research in the field of general fluid dynamics, sailboat keels, sails, aircraft wings, airfoils, ship model testing etc. in the last 100-150 years. Whatever revolutionary stuff you hope to discover will be, by 99.99% probability, just a redo of the work other people have already done (and hence a waste of your time and money) if you first don't perform a thorough research of the existing knowledge database before starting this thing.

I'm telling you this because I haven't see anything new in the research goals you have exposed so far. It is all covered by general aero/hydrodynamics theory and practice, you just need to sit down and study.

Cheers!

[Perm Stress]

Besides scale as such, absolute size of model is the most important factor.
In „Scale effects in sailing yacht hydrodinamic testing“ Transactions, SNAME New York 1974, they draw a conclusion, that 15 ft of waterline is the smallest size to get meaningfull results for sailboat model testing.

Yes, you will find that many sailboat testing is made with significantly smaller models (Delftship Series are done with 1.5-2m LWL, for example), BUT:
With smaller models they are ONLY interested in properties of HULL. Keel and rudder are only here to modify flow at their contact with hull. Appendage drag is calculated separately, deducted from the full drag to get "stripped hull" resistance, and there is an end to keel and rudder of this small scale and size. Only "Stripped Hull" data is further processed .

Simple calculations quickly show, that Reynolds number decrease dramatically with scale, because both size and speed of keel trough the water are decreasing.


Here is a calculation for 1:10 boat, with keel width 0.5m, for same scale speed=same Froude Number:

M= 10
__________Full Size______ Model
LWL=__________7______0,7 m
Froude=________0,3______0,3
V=____________2,4860____0,7861 m/s
Width of keel=__0,51______ 0,051 m
nu(20°C)= _1,004E-06____1,004E-06 m2/s
Re=________1262819_____39934

We get only ~40000 of Re for model keel. At these Reynolds, water act more like oil or syrup...

At Reynolds Number 100'000 and below, behavior of foils is not at all similar to that at 1000'000 and above. (try google "Low Reynolds", you will find enough data to confirm it) So, if model foils work at Re<100'000, there will be NO meaningful results from "simple" testing/ free sailing to extrapolate to full size.
_____________________________

So, for simple (for direct extrapolation to full size without all the tricks towing tank guys live from )research, the only advice is :
As long as possible, preferably not less than 15ft/4.7m minimum LWL.

[gonzo]

I think that making models and testing them can be a great educational tool. Books give you a lot of information, but should not prevent you from trying. The Wright brothers where bicycle mechanics, so they invented the wind tunnels to test wings. They couldn't do the math.

[gonzo]

That makes no sense that 15' is the minimum waterline length. Does that mean that models for smaller boats are full size or bigger?

[Perm Stress]

As to bilge keels, Reinke use asymmetric keels on his designs.

http://www.reinke-yacht.de/

look ASY. No much data, tough .

[Perm Stress]

Originally Posted by gonzo

That makes no sense that 15' is the minimum waterline length. Does that mean that models for smaller boats are full size or bigger?

For boats with full size less than 15ft model testing is not often applied... .
Boats so small are usually small and cheap enough to experiment full size.

Check any source on Low Reynolds data for foils. You will quickly find that Re~100'000 is a boundary between areas of radically different flows and properties of foils.

For model, it is not the LWL so important, important is to keep Re of foils close enough to 100'000 at speeds you are interested in. The wider the foils relative to LWL - the smaller the LWL of model could be.

What I post here on the subject is a result of investigation on why 0.7m LWL sailboat scale model performed faaaaaar below expectations.
Too low Reynolds on foils was one of the reasons.

[fcfc]

Originally Posted by gonzo

That makes no sense that 15' is the minimum waterline length. Does that mean that models for smaller boats are full size or bigger?

See http://www.finot.com/ecrits/recherch..._WEB_Finot.htm

(sorry french only).

The tank testing of the open 60 was scale 1:4.5 hull length 4m (13.5ft).

The dinghy testing was full scale (1:1) tank testing model = same size as the real dinghy. No information given on test price.

[gonzo]

I agree that the bigger the model, the more accurate the data. Errors also increase to the cube of the size. However, it is possible to use smaller models in side to side comparisons when you are not looking for absolute numbers. Turbulent flow is the main problem with models that are too small. Small protrusions on the leading edge have been used successfully.

[Perm Stress]

Originally Posted by gonzo

I agree that the bigger the model, the more accurate the data. Errors also increase to the cube of the size. However, it is possible to use smaller models in side to side comparisons when you are not looking for absolute numbers. Turbulent flow is the main problem with models that are too small. Small protrusions on the leading edge have been used successfully.

For hulls, turbulators proved to be acceptable solution.
For foils, however, the problem is big enough Reynolds number, based on length of chord; if you know any reference on use of turbulators to overcome low Re problem for foils, please post it.

[MastMonkey]

I finally got a chance to read some of the prior threads on twin keels and twin keel research. There is lots of excellent reading. I have to disagree that research already exist. My searching here and of the limited databases I have access to shows a lack of applied research into twin keel design. What research there is has been applied to specific hulls or one off boats. But there are manufacturers using twin keels so some studies and data must have been gathered, though I can't find it.

Thank you all for the detailed technical answers you have provided. It gives me an excellent start.

Though the keel/hull interaction seems a vita area for study with twin keel arrangements, can useful data be gathered by tested the keels independently of the hull or with only a portion of the hull?

I am not trying to make some major breakthrough, only trying to get a better understanding.

Failing that, maybe I should start a test tank business, since it seems to costs so much to use. I can use my profits to make 15' models and test them for free.

[Eralnd44]

Perm

Do not fear the wrong words of others to this list. They are coming from big education in topic and look for to criticize the one who makes not the knowing of the matter. The curious you have on this is good, ignore the person who makes it light. I see that they make no new place to understand topic and make simple wrong words for you. Good person makes the leadership. Weak person makes for the critic.

Give over to person who works to the care.

My so sorry, as to not perm and to mast monkey then as to say in last post

[frasco]

Bumpity bump from a boat designing n00b. Didn't feel like starting a new thread but I have some questions.

I'm considering building a 1 meter (~40") model of a 4 meter (~13') sailboat displacement hull. For starters, I'm interested in experimenting with keel/rudder/mast size/form/placement options on the model. I assume the balance between these should pretty much be okay (scalable) but how about lifting properties of the keel & rudder? Do foils (combined with hull form) behave the same downscaled, can I trust what the model tells me?

[Perm Stress]

Originally Posted by frasco

Bumpity bump from a boat designing n00b. Didn't feel like starting a new thread but I have some questions.

I'm considering building a 1 meter (~40") model of a 4 meter (~13') sailboat displacement hull. For starters, I'm interested in experimenting with keel/rudder/mast size/form/placement options on the model. I assume the balance between these should pretty much be okay (scalable) but how about lifting properties of the keel & rudder? Do foils (combined with hull form) behave the same downscaled, can I trust what the model tells me?

Balance will be qualitatively scalable, i.e. do not expect results accurate to to last % and applicable to full scale without fiddling/interpretation .

Foils will have less lift per area, smaller stall angle and more drag as a consequence of MUCH lower Reynolds Number.
In my little games with models, confirmed with analysis of sailing models made by other guys/companies, it proved necessary to make model foils much bigger in area than scaling would suggest.

[MalSmith]

Originally Posted by Perm Stress

Balance will be qualitatively scalable, i.e. do not expect results accurate to to last % and applicable to full scale without fiddling/interpretation .

Foils will have less lift per area, smaller stall angle and more drag as a consequence of MUCH lower Reynolds Number.
In my little games with models, confirmed with analysis of sailing models made by other guys/companies, it proved necessary to make model foils much bigger in area than scaling would suggest.

When I tested a 1:5 model of a 6m sailing multihull, I did indeed scale the foils up (relatively) for the reasons you mention. From memory I scaled them up by 20%, but I can't remember now how I arrived at that. It might have been a guess. In any case, I had no great expectation that the results would have an accurate correlation with the full size boat. In any small scale model testing, you do need to be aware of the pitfalls and make some allowances for errors.