wardd
09-27-2009, 12:05 PM
how do you scale weights for a hull model to test
thinking of making papermache hull models to test
thinking of making papermache hull models to test
View Full Version : hull model testing
hoytedow
09-27-2009, 02:30 PM
Weight of model should be directly proportional to volume of model, as compared to full size. In other words if hull volume is 1/10th that of full scale, weight of model should be 10% of full scale estimate.
I'm not a pro, so someone tell me if I am on the wrong track.
I'm not a pro, so someone tell me if I am on the wrong track.
MikeJohns
09-27-2009, 05:58 PM
Weight of model should be directly proportional to volume of model, as compared to full size. In other words if hull volume is 1/10th that of full scale, weight of model should be 10% of full scale estimate.
I'm not a pro, so someone tell me if I am on the wrong track.
Yes the displacement should be scaled by the ratio of the volumes which is a simple cubic relationship to the scale. For exmple if your scale is one tenth then (D of model) = (D Ship) / (10^3) Replace the 10 by whatever scale factor you used.
I'm not a pro, so someone tell me if I am on the wrong track.
Yes the displacement should be scaled by the ratio of the volumes which is a simple cubic relationship to the scale. For exmple if your scale is one tenth then (D of model) = (D Ship) / (10^3) Replace the 10 by whatever scale factor you used.
tom28571
09-27-2009, 08:05 PM
Correct.
ancient kayaker
09-29-2009, 12:09 AM
It is also important to get the CoG location correct.
daiquiri
09-29-2009, 04:58 AM
In addition to the above, if you plan to test the dynamic behaviour of the boat you also need to position the weights in such a way that moments of inertia are scaled properly.
See this thread:
http://www.boatdesign.net/forums/boat-design/scale-models-25246.html
See this thread:
http://www.boatdesign.net/forums/boat-design/scale-models-25246.html
Leo Lazauskas
10-31-2009, 06:33 AM
how do you scale weights for a hull model to test
thinking of making papermache hull models to test
Are you trying to estimate the drag of the full-size vessel from the model tests? If so, you will need to scale the resistance components (i.e. skin-friction and wave resistance) separately.
If the models are too small, you might have problems with surface tension and the viscosity of the water you use.
Good luck!
Leo.
thinking of making papermache hull models to test
Are you trying to estimate the drag of the full-size vessel from the model tests? If so, you will need to scale the resistance components (i.e. skin-friction and wave resistance) separately.
If the models are too small, you might have problems with surface tension and the viscosity of the water you use.
Good luck!
Leo.
MikeJohns
10-31-2009, 07:16 AM
...
If the models are too small, you might have problems with surface tension and the viscosity of the water you use....
Leo
I was just wondering if you had experience in testing small models and just how small were you able to go before this became a problem?
I have just been testing three 1.2m models which is the smallest I've ever done and the results are excellent. But it would be nice to go significantly smaller. Has anyone tested a series of standard hulls to try and see where it all falls apart?
cheers
If the models are too small, you might have problems with surface tension and the viscosity of the water you use....
Leo
I was just wondering if you had experience in testing small models and just how small were you able to go before this became a problem?
I have just been testing three 1.2m models which is the smallest I've ever done and the results are excellent. But it would be nice to go significantly smaller. Has anyone tested a series of standard hulls to try and see where it all falls apart?
cheers
ancient kayaker
10-31-2009, 09:52 AM
Are you trying to estimate the drag of the full-size vessel from the model tests? If so, you will need to scale the resistance components (i.e. skin-friction and wave resistance) separately ...
Is it practical to determine the skin-friction separately by performing a fully immersed (submerged) test then adjusting the surface results to isolate the wave resistance, since they scale differently?
If so, then the effect of the support structure can be eliminated from a submerged test by weighting a model to float stem up and releasing it at the bottom of a tall tank, noting it's speed at different bouyancies. The effect of the submerged deck can be removed by cutting off the top of model at the waterline and attaching a duplicate to it as a mirror image.
Is it practical to determine the skin-friction separately by performing a fully immersed (submerged) test then adjusting the surface results to isolate the wave resistance, since they scale differently?
If so, then the effect of the support structure can be eliminated from a submerged test by weighting a model to float stem up and releasing it at the bottom of a tall tank, noting it's speed at different bouyancies. The effect of the submerged deck can be removed by cutting off the top of model at the waterline and attaching a duplicate to it as a mirror image.
Ad Hoc
10-31-2009, 10:02 AM
Mike
We have used 800mm models very successfully. Have been validated against "proper full blown" models, ie 2m, of the same hull form, from major test taks, such as MARIN and BMT. Excellent correlation. It is all about the methodology and understanding where are the systematic errors.
A.K.
No...the skin friction is just a function of the wetted surface area.
We have used 800mm models very successfully. Have been validated against "proper full blown" models, ie 2m, of the same hull form, from major test taks, such as MARIN and BMT. Excellent correlation. It is all about the methodology and understanding where are the systematic errors.
A.K.
No...the skin friction is just a function of the wetted surface area.
marshmat
10-31-2009, 10:15 AM
As long as you're using a sound methodology and are aware of all your sources of error, even a relatively small model can provide reasonable results.
If I had the space and money, I'd prefer to work with ~ 2m models, but as it is I have to make do with computer simulations and the occasional 0.8 to 1.5 m model.
An example of the difficulties encountered with smaller models is the dilemma of matching both Froude number and Reynolds number.... the former being critical to understanding wave drag, the latter being they key to figuring out skin friction. They scale at different rates- matching a model's speed to the actual prototype's speed by Froude number results in the model's Reynolds number, at least in the forward sections, being too low. As you go smaller still, the Weber number (surface tension) starts to come into play.
It's usually possible to at least partially account for the effects of such phenomena, but to do so, you have to know what you're looking for and how scaling issues are likely to manifest themselves. Of course, a useful qualitative result can still be obtained without worrying about every last detail, but to scale speeds and forces accurately requires a fair bit of physics and math.
If I had the space and money, I'd prefer to work with ~ 2m models, but as it is I have to make do with computer simulations and the occasional 0.8 to 1.5 m model.
An example of the difficulties encountered with smaller models is the dilemma of matching both Froude number and Reynolds number.... the former being critical to understanding wave drag, the latter being they key to figuring out skin friction. They scale at different rates- matching a model's speed to the actual prototype's speed by Froude number results in the model's Reynolds number, at least in the forward sections, being too low. As you go smaller still, the Weber number (surface tension) starts to come into play.
It's usually possible to at least partially account for the effects of such phenomena, but to do so, you have to know what you're looking for and how scaling issues are likely to manifest themselves. Of course, a useful qualitative result can still be obtained without worrying about every last detail, but to scale speeds and forces accurately requires a fair bit of physics and math.
Crag Cay.
10-31-2009, 10:46 AM
The operation of a towing tank is a highly competitive commercial field. The ability to use smaller models to give a stipulated level of accuracy is one of the factors that can make a particular facility competitive. It lowers costs and increases the scope of work that can be undertaken. Detailed benchmarking data is unlikely to be in the public domain.
Leo Lazauskas
10-31-2009, 11:24 AM
Leo
I was just wondering if you had experience in testing small models and just how small were you able to go before this became a problem?
I have just been testing three 1.2m models which is the smallest I've ever done and the results are excellent. But it would be nice to go significantly smaller. Has anyone tested a series of standard hulls to try and see where it all falls apart?
cheers
I personally don't do any experimental work myself, however I do
have a lot of data from a variety of tanks.
In some ways it depends on what you are testing and how far you
will be extrapolating from models to full-size.
Preliminary results from a currently running ITTC project,
"Uncertainty analysis for towing tank measurements", suggests
that there is more scatter in the results for a 3.0m destroyer hull
than for its 5.72m geosim.
Those preliminary results also show significant differences between
results from different tanks.
Results I have seen for many tests on simple Wigley hulls vary
considerably. For example Insel's experiments on a 1.5m hull don't
square up well with experiments with 3.048m hulls.
Similarly, there is a lot of scatter in some of Couser's results with
1.6m NPL series hulls.
Effects on (theoretical) wave resistance suggest that models should
be longer than about 1.5m. A classic work on this is:
Webster, William C.,
"The effect of surface tension on wave resistance"
Report No. NA-66-6
College of Engineering,
Uni California, Berkeley, July 1966.
(Sorry, but I only have a bound original of this.)
A meniscus of 1.5mm might be significant on a small slender (i.e.
small draft) hull, but of no import on the wetted area of a full-size
ship. Experimental results for hulls that create a lot of spray might not
be approximated well at model scale.
Large transom sterns are another potential source of difficulties
for experiments (and theory!). Splash, spray and wave-breaking
effects might not scale conveniently.
As for testing a systematic series of a small hulls...
L.J.Doctors supervised a set of experiments on a series of mathematical
hulls with systematically varying amounts of parallel middlebody and
transom size. Not quite what you were after, but there are some
interesting scaling issues. Some of the models were about 0.8m or smaller.
I've attached some of Lawry's papers. Personally, I don't agree with the
transom hollow approach that he and Couser et al use. I don't believe
that a hollow behind the stern (and ventilated to atmospheric pressure)
can create waves as if it was an extension of the hull.
Another difficulty that I see is that some effects that are
apparent at small scale might not be important at full size,
and vice versa. Therefore there can be some cancellations
and reinforcements that might not be easy to appreciate.
Some tests at lengths of 1.0m might correlate well with tests
at 2m or more, but not necessarily to full-size.
For example, in towing tanks, the eddy viscosity might be quite
different to that experienced by vessels at sea. Eddy viscosity damps
out high-frequency waves, and can vary by up to 5 or 6 orders at sea.
On the other hand, surface tension effects won't be important on
real vessels.
How ambient eddy viscosity affects skin-friction is another matter
altogether and I don't know if anyone has done much on that.
And now a question for you...
I have been working on the resistance of ships travelling
through a surface layer of ice slush. Have you seen any
experimental results for that kind of thing, or for ships travelling
through oil-slicks?
All the best,
Leo.
I was just wondering if you had experience in testing small models and just how small were you able to go before this became a problem?
I have just been testing three 1.2m models which is the smallest I've ever done and the results are excellent. But it would be nice to go significantly smaller. Has anyone tested a series of standard hulls to try and see where it all falls apart?
cheers
I personally don't do any experimental work myself, however I do
have a lot of data from a variety of tanks.
In some ways it depends on what you are testing and how far you
will be extrapolating from models to full-size.
Preliminary results from a currently running ITTC project,
"Uncertainty analysis for towing tank measurements", suggests
that there is more scatter in the results for a 3.0m destroyer hull
than for its 5.72m geosim.
Those preliminary results also show significant differences between
results from different tanks.
Results I have seen for many tests on simple Wigley hulls vary
considerably. For example Insel's experiments on a 1.5m hull don't
square up well with experiments with 3.048m hulls.
Similarly, there is a lot of scatter in some of Couser's results with
1.6m NPL series hulls.
Effects on (theoretical) wave resistance suggest that models should
be longer than about 1.5m. A classic work on this is:
Webster, William C.,
"The effect of surface tension on wave resistance"
Report No. NA-66-6
College of Engineering,
Uni California, Berkeley, July 1966.
(Sorry, but I only have a bound original of this.)
A meniscus of 1.5mm might be significant on a small slender (i.e.
small draft) hull, but of no import on the wetted area of a full-size
ship. Experimental results for hulls that create a lot of spray might not
be approximated well at model scale.
Large transom sterns are another potential source of difficulties
for experiments (and theory!). Splash, spray and wave-breaking
effects might not scale conveniently.
As for testing a systematic series of a small hulls...
L.J.Doctors supervised a set of experiments on a series of mathematical
hulls with systematically varying amounts of parallel middlebody and
transom size. Not quite what you were after, but there are some
interesting scaling issues. Some of the models were about 0.8m or smaller.
I've attached some of Lawry's papers. Personally, I don't agree with the
transom hollow approach that he and Couser et al use. I don't believe
that a hollow behind the stern (and ventilated to atmospheric pressure)
can create waves as if it was an extension of the hull.
Another difficulty that I see is that some effects that are
apparent at small scale might not be important at full size,
and vice versa. Therefore there can be some cancellations
and reinforcements that might not be easy to appreciate.
Some tests at lengths of 1.0m might correlate well with tests
at 2m or more, but not necessarily to full-size.
For example, in towing tanks, the eddy viscosity might be quite
different to that experienced by vessels at sea. Eddy viscosity damps
out high-frequency waves, and can vary by up to 5 or 6 orders at sea.
On the other hand, surface tension effects won't be important on
real vessels.
How ambient eddy viscosity affects skin-friction is another matter
altogether and I don't know if anyone has done much on that.
And now a question for you...
I have been working on the resistance of ships travelling
through a surface layer of ice slush. Have you seen any
experimental results for that kind of thing, or for ships travelling
through oil-slicks?
All the best,
Leo.
yipster
10-31-2009, 11:28 AM
my pro memory note say's: linear dimensions only scale
areas be reduced by the square root of the scale
volumes and weights be reduced by the cube
velocity's scale by the square root of scaling factor
areas be reduced by the square root of the scale
volumes and weights be reduced by the cube
velocity's scale by the square root of scaling factor
Leo Lazauskas
10-31-2009, 12:16 PM
Leo
Has anyone tested a series of standard hulls to try and see where it all falls apart?
Just one more important consideration for very small models...
The size, spacing and location of boundary-layer tripping devices.
Leo.
Has anyone tested a series of standard hulls to try and see where it all falls apart?
Just one more important consideration for very small models...
The size, spacing and location of boundary-layer tripping devices.
Leo.
wardd
10-31-2009, 12:37 PM
i was thinking weights and balance and loading and to get an idea of how it would behave , rolling etc?
Leo Lazauskas
10-31-2009, 03:02 PM
Is it practical to determine the skin-friction separately by performing a fully immersed (submerged) test then adjusting the surface results to isolate the wave resistance, since they scale differently?
If so, then the effect of the support structure can be eliminated from a submerged test by weighting a model to float stem up and releasing it at the bottom of a tall tank, noting it's speed at different bouyancies. The effect of the submerged deck can be removed by cutting off the top of model at the waterline and attaching a duplicate to it as a mirror image.
It's probably not a good method because (among others):
1. You wont get the same type of separation near the stern, and you might get unusual flow separation at the bow. Hence there might be some very odd form drag effects.
2. Skin-friction depends mostly on the wetted surface area.
3. There are some interactions between the waves on the hull and the boundary layer for surface-piercing vessels, so skin-friction is not exactly proportional to surface area. Remember: it's Froude's "Hypothesis", not Froude's "Law
" :)
The idea of a double body is used in some analytical treatments (albeit in a different way), so your proposal is not a completely silly idea.
Leo.
If so, then the effect of the support structure can be eliminated from a submerged test by weighting a model to float stem up and releasing it at the bottom of a tall tank, noting it's speed at different bouyancies. The effect of the submerged deck can be removed by cutting off the top of model at the waterline and attaching a duplicate to it as a mirror image.
It's probably not a good method because (among others):
1. You wont get the same type of separation near the stern, and you might get unusual flow separation at the bow. Hence there might be some very odd form drag effects.
2. Skin-friction depends mostly on the wetted surface area.
3. There are some interactions between the waves on the hull and the boundary layer for surface-piercing vessels, so skin-friction is not exactly proportional to surface area. Remember: it's Froude's "Hypothesis", not Froude's "Law
" :)
The idea of a double body is used in some analytical treatments (albeit in a different way), so your proposal is not a completely silly idea.
Leo.
View Full Version : hull model testing
Legal •
Boat Design Forums •
Gallery •
Directory •
Boat Design & Boatbuilding Bookstore •
Powerboat Ads •
Sailboat Ads
powered by vBulletin which is copyright ©2000-2008, Jelsoft Enterprises Ltd.
Web Site Design and Content Copyright ©1999-2008 Boat Design Net
powered by vBulletin which is copyright ©2000-2008, Jelsoft Enterprises Ltd.
Web Site Design and Content Copyright ©1999-2008 Boat Design Net