Scale Model testing

[Yachti]

Hi Guys. I am in the progress of testing a 1/30th scale model of a planning hull. I plan to sclae both the weigh and speed of the hull. How i the best method of doing this? when i use reynolds number to scale the speed, i get a number similar to 750 m/s??? also the weight seems rathe rhigh aswel.

Many thanks

[messabout]

One thirtieth scale is much too small for meaningful observations. For example there is the weight scaling problem. Use the scale factor cubed...30 x 30 x 30 = 27000..... That is the divisor to use and the weight of the full sized boat is used as the dividend....Example: suppose the full sized boat is to weigh 2700 pounds. 2700/27000= 0.1.....your model will need to displace 0.1 pounds to simulate the 2700 pound boat. That is not a practical weight for the model. Now consider using something on the order of one eighth scale for the model. Scale factor 8 cubed = 512...and your 2700 pound big boat is simulated by 2700/512 = 5.27 pounds. which is a more realistic number. If you are thinking of a small planing boat that might weigh 800 pounds then the scale weight problem is even more serious.....800/27000 = 0.0296 pounds. If you are working in the SI system then kilogram weights work just as well. Make the model bigger. You will get much better results.

[Evaldasnar]

Hi,

In model tests its important to Have the same Froude Number Fr=v/sqrt(g*L). The same time its not impossible heave Same Froude and Reynolds numbers .

The model characteristics:
Geometrical scale - k
Areas - k^2
Volumes, masses, forces - k^3

This is main scales, its much more. Say, if you need it.


Sorry for my English

[kach22i]

For hovercraft and SES 1/6th scale applies as best. I don't know about boats.

[tom28571]

Most who have tested towing models don't recommend anything less than 4' (1.2M) length, whatever the scale. Performance scaling factors become even more suspect in small sizes.

[LyndonJ]

There are professionals on this forum who have used models down to 800mm successfully. The crux is accuracy and careful recording of all the factors that affect resistance.
Very important to make sure the flow has tripped !

[u4ea32]

I have read many studies over many years by many professionals that suggest that scale effects require large models, or the results can be simply WRONG.

A famous example was the Chance designed Mariner 12 meter.

Some recent studies have found that about 20 feet is the minimum for semi planing and planing models.

In order to get past the heavy biases that anyone develops following hard work, its not effective to simply ask people what they found useful.

People make huge mistakes, and if they put enough effort in, they will deny the mistake to the end. Religions are a great example of this.

I suggest instead using an independent approach to decide what the proper scale should be. The independent approach is to see at what point competition boats start looking the same.

For example, the outer hulls of BMW Oracle look very similar to the hulls of an A class catamaran. Therefore, 17 feet is sufficient for long narrow hulls that operate from SL ratios of 1 to 4.

Speed boats keep looking quite different until about 30 feet: a 20' Allison looks nothing like a 42' Fountain. So even a 20 foot scale model would have limited utility in high speed regimes.

On keel boats, its hard to tell as there are so few open classes that allow a large number of designers a free hand to explore the design space, and the performance of all such boats is so dependent upon the sailors and weather routing. For example, are chines fast or slow? Some people think they are fast because of the JK boats that so dominated the last Volvo race. But there were others in that race by Farr that were off the pace -- so do chines work or not? And there are several other boats who had major surgery to REMOVE those JK chines, and the boat after change became dramatically faster. So whose tank tests and CFD worked, and whose did not? Hard to tell. Or look at the America's cup boats. Some teams went to 30 foot models for those 80 foot monohulls. The boats that really did well actually used full scale -- they built multiple boats, and then they replaced large portions of the boat several times.

In other words, it looks like scale models have very limited utility if one is really concerned about precision. It depends on how much error you are willing to accept. The problem is the error is greater than the effect of most major design choices you would want to experiment with, including deadrise, beam, and so on.

Luckily, there is a lot of data available, and a lot of reduction has been done, so you can do a decent job of modeling mathematically. And you can always modify the eventual boat.

[LyndonJ]

Originally Posted by u4ea32

I have read many studies over many years by many professionals that suggest that scale effects require large models, or the results can be simply WRONG.....

Yes but how wrong is the important question isn't it ? And planing and semi planing are in a different arena to displacment craft.

2% wrong will land you in court for a racing boat but for a ocmmercial vessel it's quite accurate for indicating powering.

The tank testing facilities around the globe often refine hullforms on 1.8m models. If you actually get involved with some of these tests it's quite illuminating to see just how much you can garner from it.

For example wave making and the wave slice alone is invaluable, and you can get that from very small scales more accurately than with larger models.

So depends what you are testing and what you are refining.

The hard and fast rule is to always anchor your test results to real world data anyway, once you get away from this or you are looking for that small % edge then maybe it's not so useful.

[yipster]

http://www.youtube.com/watch?v=mNVVTQaXSvc
ok a pull prop but its still amazing

[Capt JZ]

You guys are saying really good stuff. With so many variables, I like interpolation within the tank test data and the real (full scale) world data; the different qualities do scale differently.