University project: Multihull wave making drag

[leegrace]

Hello,

I have been browsing the forums and found some very useful and interesting information from a collaboration of great minds...

I am currently in my final year at University and I am embarking on a research project looking into reducing the wave making drag of multihulls by optimising hull spacing.

I am at the very early stages of my research but any information or help along the way would be greatly appreciated.

The rough plan for my project is to do a literature review and research on papers and information already in existance. This will be followed up by some testing on a computational fluid dynamics (CFX) package using some test models created in maxsurf. Hopefully this will lead to some interesting results and allow me to make a scale model to test in the universities flow tank.

I originally wanted to get an understanding of how the spacing of catamaran hulls affect wave making drag, i.e. constructive/destructive wave patterns being the resultant. This has led me to this investigation with my outcome hopefully being able to produce some design parameters allowing the amateur designer to use my database to help initial planning, sizing and spacing of hulls. I am well aware that much of the wave making drag can be reduced by optimising the shape and form of the hulls in the first place but I beleive there to be room for improvement into the understanding of hull spacing.

So, hopefully anyone with any input will be willing to contribute to this discussion and help me and others to understand this aspect of multihull design a little more.

Thanks
Lee

[RANCHI OTTO]

Interesting paper :
Design and application of modern high-speed catamaran
Sname publication

Several tank test results with influence of clearance btw. hulls

[Leo Lazauskas]

Hello,

I have been browsing the forums and found some very useful and interesting information from a collaboration of great minds...

I am currently in my final year at University and I am embarking on a research project looking into reducing the wave making drag of multihulls by optimising hull spacing.

I am at the very early stages of my research but any information or help along the way would be greatly appreciated.

The rough plan for my project is to do a literature review and research on papers and information already in existance. This will be followed up by some testing on a computational fluid dynamics (CFX) package using some test models created in maxsurf. Hopefully this will lead to some interesting results and allow me to make a scale model to test in the universities flow tank.

I originally wanted to get an understanding of how the spacing of catamaran hulls affect wave making drag, i.e. constructive/destructive wave patterns being the resultant. This has led me to this investigation with my outcome hopefully being able to produce some design parameters allowing the amateur designer to use my database to help initial planning, sizing and spacing of hulls. I am well aware that much of the wave making drag can be reduced by optimising the shape and form of the hulls in the first place but I beleive there to be room for improvement into the understanding of hull spacing.

So, hopefully anyone with any input will be willing to contribute to this discussion and help me and others to understand this aspect of multihull design a little more.

Thanks
Lee

Have a look at:
http://www.cyberiad.net/multihulls.htm

and in particular the paper titled "Optimum Spacing of a Family of Multihulls" by E.O.Tuck and L. Lazauskas.

Good luck with your studies!
Leo.

[Leo Lazauskas]

I understand that Mitchell is for thin ships, Do you have a feel for how "thick" you can go and still get good results?

I have found that Michell (pronounced Mitchell, but not spelled that way) theory works quite well for hulls with L/B greater than about 6. However, others have found good agreement with experiments for hulls down to L/B=4. I guess it depends on how much you believe some experiments, and how much you trust theory. :)

One good thing about Michell (and many other imple physics-based) codes is that you can usually trust the ranking of the predicted performance of the hulls. Thus if Hull A has a lower predicted drag than Hull B, that is usually also what experiments will show. The predicted drag may, however, not be a particularly good estimate of the actual drag.

Leo.

[leegrace]

Thanks for your quick responses guys.

Leo: Yours was one of the first papers I came across and has given me much to think about.

Ranchi: Thanks also, I have not come across that paper yet but I will endeavour to find it.

I will keep this discussion up to date with my findings and results and no doubt asking many questions.

Lee

[masrapido]

Nice paper "Optimum spacing for a family of multihulls". Very informative reading. Made me download Michell and give up on it as soon as I opened it. I wonder what would be the results of spacing if hull design were taken into account. In reality spacing alone doesn't play a role. It only plays a role coupled with proper hull design. Is there a paper that covers such a situation? Maybe by you Leo?

(how on earth one uses Michell...?)

[Leo Lazauskas]

Nice paper "Optimum spacing for a family of multihulls". Very informative reading. Made me download Michell and give up on it as soon as I opened it. I wonder what would be the results of spacing if hull design were taken into account. In reality spacing alone doesn't play a role. It only plays a role coupled with proper hull design. Is there a paper that covers such a situation? Maybe by you Leo?

(how on earth one uses Michell...?)

It is not an easy program for beginners. You must read the manual and then work through the examples.

Michlet can be used to simultaneously optimise the spacing and hull shape for multihulls with up to 5 demihulls, and they do not necessarily have to be symmetrically arranged (although the demihulls themselves are assumed to be symmetric with respect to their own centreplanes).

There is a preprint of a thesis that looks at optimisation of monohulls, cats, trimarans, Surface Effect Ships, and hovercraft at:

www.cyberiad.net/hybrid.htm

The vessels in that work, however, are optimised at 50 knots and 75 knots, and subject to a variety of practical constraints.

There is some other work in:
Lazauskas, L. and Tuck, E.O., Low drag multihulls for sporting,
commercial and military applications, FAST97, pp. 647-652.

Sorry, but I don't have an electronic copy. This paper looks at large (3500 tonne) monohulls, cats and tris. The shape of the hulls was kept constant, although length, beam, draft, and hull spacing was allowed to vary.

There are many papers given at the FAST series of conferences that you could start with. A good university (or other) library should be able to get them in for you.

If you are really keen on this subject, be prepared to put in a lot of work to understand it. As you might have gathered from the Schiffstechnik paper, the free wave spectrum is the key to understanding wave cancellation. Michlet will calculate this for you, but be prepared to put in a few hours.

Have fun!
Leo.

[masrapido]

xxnxn

[masrapido]

I thought that would happen...
:cool:
Here's the real message:

Leo,

I discovered an excellent manual and tutorials in spanish written by Ramon Heernandez Ortega, and it turned out to be a 30 minutes learning curve. On Michlet's functionalities, commands and output. (apologies for calling it Michell...).

Reading one of your papers the other day I got a bit puzzled with parabolic and elyptical shapes for hulls. Am I wrong in having impression that you are suggesting these to be most appropriate for transverse wave cancelling? (language barrier...reading can be challenging, especially with such heavy content)

[Leo Lazauskas]

Reading one of your papers the other day I got a bit puzzled with parabolic and elyptical shapes for hulls. Am I wrong in having impression that you are suggesting these to be most appropriate for transverse wave cancelling? (language barrier...reading can be challenging, especially with such heavy content)

Yes, you are probably wrong. :) What paper are you referring to? In some papers we fixed the hull shape because we were only interested in trying to find the optimum length. In other papers we let the shape vary. In general, hull shape is not important for transverse wave cancellation: hull length is a far more important parameter.

Have fun!
Leo.

[tspeer]

I think Leo picked the parabolic, elliptical and rectangular hullforms because they were convex and representative of both real-life hull forms and analytical studies in the literature. For example, parabolic sections result in deadrise with a rounded bilge. Elliptical sections produce rounded bottoms, and rectangular sections produce cylindrical struts. Between the sections, rocker, and waterlines, these gave him 27 possible variations in hull shape for his optimization. And they weren't subject to strange wiggles like some hull shapes that are optimized for minimum wave drag at specific conditions (hence the need to be convex).

But there are important things that these simple hull shapes couldn't capture, like positioning the longitudinal center of buoyancy. So for his thesis (and I can't believe you've only recently received your MS, Leo!), he used a richer variety of hull forms that included a forebody, mid section, and after body.

The problem is how to define a hull with a tractable number of parameters. There are many possible families of hull shapes that could be used to minimize the resistance within that family of hull forms and still produce a practical shape. For example, here is an approach (http://www.basiliscus.com/CaseStudy/geometry.html) that starts with a sectional area distribution defined by the length, volume, center of buoyancy, and prismatic coefficient, then uses a small number of parameters (depth/beam ratio, deadrise, height and slope of the bottom/topsides join, and sharpness of the bilge) to define the section shape at bow, stern and midships.

I think you should pick a family of hull shapes that are appropriate for the type of problem you're solving. Then use Michlet/Godzilla to find the optimum member of your family of hull forms.

[Leo Lazauskas]

I think Leo picked the parabolic, elliptical and rectangular hullforms because they were convex and representative of both real-life hull forms and analytical studies in the literature. For example, parabolic sections result in deadrise with a rounded bilge. Elliptical sections produce rounded bottoms, and rectangular sections produce cylindrical struts. Between the sections, rocker, and waterlines, these gave him 27 possible variations in hull shape for his optimization. And they weren't subject to strange wiggles like some hull shapes that are optimized for minimum wave drag at specific conditions (hence the need to be convex).


That pretty well sums up my reasons with a few minor comments...

There are more than 27 possible variations. A hull can have shapes "between" parabolic and elliptical, for example, if 0.5 represents an elliptical section and 1.0 represents a parabolic shape, then 0.75 is a shape somewhere between the two.

It is possible to use hull shape parameters that are non-convex. For example, if parameters are greater than 1 the shape will be cusped.

Another reason is that it possible to represent an entire hull with just a few principal dimensions and some shape parameters. This allows me to use evolutionary algorithms for optimisation. Without such a representation it would be difficult to fit all the hulls into PC memory, or the code would run impractically slowly.


So for his thesis (and I can't believe you've only recently received your MS, Leo!)...

My wife and the prof eventually ground me down to get some formal academic qualifications. I held out for 12 years but I'm hooked now and 6 months into a PhD. :)

Leo Lazauskas PhD (Pending).

[masrapido]

I think Leo picked the parabolic, elliptical and rectangular hullforms because they were convex and representative of both real-life hull forms and analytical studies in the literature.

It is possible, but it was the drag that was optimised in the article, not the waves. Reading Leo's papers in such a short time, I got them mixed up. Also, I am not solving any problems yet, that was the guy who started the thread. I will face technical problems once I finish visual/graphical design of my own catamaran.

Would it be okay to send a draft to either of you for technical comments once I come up with a shape I like?

[Leo Lazauskas]

It is possible, but it was the drag that was optimised in the article, not the waves. Reading Leo's papers in such a short time, I got them mixed up. Also, I am not solving any problems yet, that was the guy who started the thread. I will face technical problems once I finish visual/graphical design of my own catamaran.

Would it be okay to send a draft to either of you for technical comments once I come up with a shape I like?

You can send it to me, but remember that I am not a very practical person. To me, boats are just mathematical objects.

Leo.

[masrapido]

Ok, thanks. Mathematics are of great interest once the visual form is defined. I know now that I can test the hull in Michlet, but a quick comment on lines/shape/drag from mathematical point of view would be helpful in understanding the hulls limitations and optimum speed one could expect.

[marshmat]

In general the length will be the main factor in transverse wave cancellation... but I would note that an asymmetrical hull with a very flat inboard surface would likely not produce as much of a wave between the hulls as a symmetrical hull would. Just a thought.....
On CFD: I'm most familiar with Fluent (and not very, at that...); generally, CFD is great at comparison but poor at absolutes. That is, for two hulls, the ratio of any given parameter for one hull to the same parameter for the other will turn out very close to what it is in reality... but the actual values of those parameters will probably both be wrong by a bit. So it's great as a comparison tool, but is no substitute for a tank or a real hull in performance prediction.

[Leo Lazauskas]

In general the length will be the main factor in transverse wave cancellation... but I would note that an asymmetrical hull with a very flat inboard surface would likely not produce as much of a wave between the hulls as a symmetrical hull would. Just a thought.....


It's possible, but I've never seen any good experimental results demonstrating it. It's not clear that there is any significant resistance reduction by using flat inboard surfaces, especially after induced drag is included in the total resistance.

Prof. Ernie Tuck at the University of Adelaide recently gave a paper on this topic. The paper is at:

http://internal.maths.adelaide.edu.au/people/etuck/pdfiles/vortex04.pdf

What he found (and I have verified numerically as well) is that asymmetric demihulls might be beneficial if the demihulls are not at their optimum spacing for a particular speed. If the demihulls are located near their optimum position, it is unlikely that you can reduce the drag by using asymmetric demihulls, or by using toe-in or toe out.

Regards,
Leo.